4 The following is a consolidated list of the kernel parameters as
5 implemented by the __setup(), core_param() and module_param() macros
6 and sorted into English Dictionary order (defined as ignoring all
7 punctuation and sorting digits before letters in a case insensitive
8 manner), and with descriptions where known.
10 The kernel parses parameters from the kernel command line up to "--";
11 if it doesn't recognize a parameter and it doesn't contain a '.', the
12 parameter gets passed to init: parameters with '=' go into init's
13 environment, others are passed as command line arguments to init.
14 Everything after "--" is passed as an argument to init.
16 Module parameters can be specified in two ways: via the kernel command
17 line with a module name prefix, or via modprobe, e.g.:
19 (kernel command line) usbcore.blinkenlights=1
20 (modprobe command line) modprobe usbcore blinkenlights=1
22 Parameters for modules which are built into the kernel need to be
23 specified on the kernel command line. modprobe looks through the
24 kernel command line (/proc/cmdline) and collects module parameters
25 when it loads a module, so the kernel command line can be used for
28 Hyphens (dashes) and underscores are equivalent in parameter names, so
29 log_buf_len=1M print-fatal-signals=1
30 can also be entered as
31 log-buf-len=1M print_fatal_signals=1
33 Double-quotes can be used to protect spaces in values, e.g.:
34 param="spaces in here"
36 This document may not be entirely up to date and comprehensive. The command
37 "modinfo -p ${modulename}" shows a current list of all parameters of a loadable
38 module. Loadable modules, after being loaded into the running kernel, also
39 reveal their parameters in /sys/module/${modulename}/parameters/. Some of these
40 parameters may be changed at runtime by the command
41 "echo -n ${value} > /sys/module/${modulename}/parameters/${parm}".
43 The parameters listed below are only valid if certain kernel build options were
44 enabled and if respective hardware is present. The text in square brackets at
45 the beginning of each description states the restrictions within which a
46 parameter is applicable:
48 ACPI ACPI support is enabled.
49 AGP AGP (Accelerated Graphics Port) is enabled.
50 ALSA ALSA sound support is enabled.
51 APIC APIC support is enabled.
52 APM Advanced Power Management support is enabled.
53 ARM ARM architecture is enabled.
54 AVR32 AVR32 architecture is enabled.
55 AX25 Appropriate AX.25 support is enabled.
56 BLACKFIN Blackfin architecture is enabled.
57 CLK Common clock infrastructure is enabled.
58 CMA Contiguous Memory Area support is enabled.
59 DRM Direct Rendering Management support is enabled.
60 DYNAMIC_DEBUG Build in debug messages and enable them at runtime
61 EDD BIOS Enhanced Disk Drive Services (EDD) is enabled
62 EFI EFI Partitioning (GPT) is enabled
63 EIDE EIDE/ATAPI support is enabled.
64 EVM Extended Verification Module
65 FB The frame buffer device is enabled.
66 FTRACE Function tracing enabled.
67 GCOV GCOV profiling is enabled.
68 HW Appropriate hardware is enabled.
69 IA-64 IA-64 architecture is enabled.
70 IMA Integrity measurement architecture is enabled.
71 IOSCHED More than one I/O scheduler is enabled.
72 IP_PNP IP DHCP, BOOTP, or RARP is enabled.
73 IPV6 IPv6 support is enabled.
74 ISAPNP ISA PnP code is enabled.
75 ISDN Appropriate ISDN support is enabled.
76 JOY Appropriate joystick support is enabled.
77 KGDB Kernel debugger support is enabled.
78 KVM Kernel Virtual Machine support is enabled.
79 LIBATA Libata driver is enabled
80 LP Printer support is enabled.
81 LOOP Loopback device support is enabled.
82 M68k M68k architecture is enabled.
83 These options have more detailed description inside of
84 Documentation/m68k/kernel-options.txt.
85 MDA MDA console support is enabled.
86 MIPS MIPS architecture is enabled.
87 MOUSE Appropriate mouse support is enabled.
88 MSI Message Signaled Interrupts (PCI).
89 MTD MTD (Memory Technology Device) support is enabled.
90 NET Appropriate network support is enabled.
91 NUMA NUMA support is enabled.
92 NFS Appropriate NFS support is enabled.
93 OSS OSS sound support is enabled.
94 PV_OPS A paravirtualized kernel is enabled.
95 PARIDE The ParIDE (parallel port IDE) subsystem is enabled.
96 PARISC The PA-RISC architecture is enabled.
97 PCI PCI bus support is enabled.
98 PCIE PCI Express support is enabled.
99 PCMCIA The PCMCIA subsystem is enabled.
100 PNP Plug & Play support is enabled.
101 PPC PowerPC architecture is enabled.
102 PPT Parallel port support is enabled.
103 PS2 Appropriate PS/2 support is enabled.
104 RAM RAM disk support is enabled.
105 S390 S390 architecture is enabled.
106 SCSI Appropriate SCSI support is enabled.
107 A lot of drivers have their options described inside
108 the Documentation/scsi/ sub-directory.
109 SECURITY Different security models are enabled.
110 SELINUX SELinux support is enabled.
111 APPARMOR AppArmor support is enabled.
112 SERIAL Serial support is enabled.
113 SH SuperH architecture is enabled.
114 SMP The kernel is an SMP kernel.
115 SPARC Sparc architecture is enabled.
116 SWSUSP Software suspend (hibernation) is enabled.
117 SUSPEND System suspend states are enabled.
118 TPM TPM drivers are enabled.
119 TS Appropriate touchscreen support is enabled.
120 UMS USB Mass Storage support is enabled.
121 USB USB support is enabled.
122 USBHID USB Human Interface Device support is enabled.
123 V4L Video For Linux support is enabled.
124 VMMIO Driver for memory mapped virtio devices is enabled.
125 VGA The VGA console has been enabled.
126 VT Virtual terminal support is enabled.
127 WDT Watchdog support is enabled.
128 XT IBM PC/XT MFM hard disk support is enabled.
129 X86-32 X86-32, aka i386 architecture is enabled.
130 X86-64 X86-64 architecture is enabled.
131 More X86-64 boot options can be found in
132 Documentation/x86/x86_64/boot-options.txt .
133 X86 Either 32-bit or 64-bit x86 (same as X86-32+X86-64)
134 XEN Xen support is enabled
136 In addition, the following text indicates that the option:
138 BUGS= Relates to possible processor bugs on the said processor.
139 KNL Is a kernel start-up parameter.
140 BOOT Is a boot loader parameter.
142 Parameters denoted with BOOT are actually interpreted by the boot
143 loader, and have no meaning to the kernel directly.
144 Do not modify the syntax of boot loader parameters without extreme
145 need or coordination with <Documentation/x86/boot.txt>.
147 There are also arch-specific kernel-parameters not documented here.
148 See for example <Documentation/x86/x86_64/boot-options.txt>.
150 Note that ALL kernel parameters listed below are CASE SENSITIVE, and that
151 a trailing = on the name of any parameter states that that parameter will
152 be entered as an environment variable, whereas its absence indicates that
153 it will appear as a kernel argument readable via /proc/cmdline by programs
154 running once the system is up.
156 The number of kernel parameters is not limited, but the length of the
157 complete command line (parameters including spaces etc.) is limited to
158 a fixed number of characters. This limit depends on the architecture
159 and is between 256 and 4096 characters. It is defined in the file
160 ./include/asm/setup.h as COMMAND_LINE_SIZE.
162 Finally, the [KMG] suffix is commonly described after a number of kernel
163 parameter values. These 'K', 'M', and 'G' letters represent the _binary_
164 multipliers 'Kilo', 'Mega', and 'Giga', equalling 2^10, 2^20, and 2^30
165 bytes respectively. Such letter suffixes can also be entirely omitted.
168 acpi= [HW,ACPI,X86,ARM64]
169 Advanced Configuration and Power Interface
170 Format: { force | off | strict | noirq | rsdt |
172 force -- enable ACPI if default was off
173 off -- disable ACPI if default was on
174 noirq -- do not use ACPI for IRQ routing
175 strict -- Be less tolerant of platforms that are not
176 strictly ACPI specification compliant.
177 rsdt -- prefer RSDT over (default) XSDT
178 copy_dsdt -- copy DSDT to memory
179 For ARM64, ONLY "acpi=off" or "acpi=force" are available
181 See also Documentation/power/runtime_pm.txt, pci=noacpi
183 acpi_apic_instance= [ACPI, IOAPIC]
185 2: use 2nd APIC table, if available
186 1,0: use 1st APIC table
189 acpi_backlight= [HW,ACPI]
190 acpi_backlight=vendor
192 If set to vendor, prefer vendor specific driver
193 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
194 of the ACPI video.ko driver.
196 acpica_no_return_repair [HW, ACPI]
197 Disable AML predefined validation mechanism
198 This mechanism can repair the evaluation result to make
199 the return objects more ACPI specification compliant.
200 This option is useful for developers to identify the
201 root cause of an AML interpreter issue when the issue
202 has something to do with the repair mechanism.
204 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
205 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
207 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
208 debug output. Bits in debug_layer correspond to a
209 _COMPONENT in an ACPI source file, e.g.,
210 #define _COMPONENT ACPI_PCI_COMPONENT
211 Bits in debug_level correspond to a level in
212 ACPI_DEBUG_PRINT statements, e.g.,
213 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
214 The debug_level mask defaults to "info". See
215 Documentation/acpi/debug.txt for more information about
216 debug layers and levels.
218 Enable processor driver info messages:
219 acpi.debug_layer=0x20000000
220 Enable PCI/PCI interrupt routing info messages:
221 acpi.debug_layer=0x400000
222 Enable AML "Debug" output, i.e., stores to the Debug
223 object while interpreting AML:
224 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
225 Enable all messages related to ACPI hardware:
226 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
228 Some values produce so much output that the system is
229 unusable. The "log_buf_len" parameter may be useful
230 if you need to capture more output.
232 acpi_enforce_resources= [ACPI]
233 { strict | lax | no }
234 Check for resource conflicts between native drivers
235 and ACPI OperationRegions (SystemIO and SystemMemory
236 only). IO ports and memory declared in ACPI might be
237 used by the ACPI subsystem in arbitrary AML code and
238 can interfere with legacy drivers.
239 strict (default): access to resources claimed by ACPI
240 is denied; legacy drivers trying to access reserved
241 resources will fail to bind to device using them.
242 lax: access to resources claimed by ACPI is allowed;
243 legacy drivers trying to access reserved resources
244 will bind successfully but a warning message is logged.
245 no: ACPI OperationRegions are not marked as reserved,
246 no further checks are performed.
248 acpi_force_table_verification [HW,ACPI]
249 Enable table checksum verification during early stage.
250 By default, this is disabled due to x86 early mapping
253 acpi_irq_balance [HW,ACPI]
254 ACPI will balance active IRQs
257 acpi_irq_nobalance [HW,ACPI]
258 ACPI will not move active IRQs (default)
261 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
262 Format: <irq>,<irq>...
264 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
266 Format: <irq>,<irq>...
268 acpi_no_auto_serialize [HW,ACPI]
269 Disable auto-serialization of AML methods
270 AML control methods that contain the opcodes to create
271 named objects will be marked as "Serialized" by the
272 auto-serialization feature.
273 This feature is enabled by default.
274 This option allows to turn off the feature.
276 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
279 acpi_no_static_ssdt [HW,ACPI]
280 Disable installation of static SSDTs at early boot time
281 By default, SSDTs contained in the RSDT/XSDT will be
282 installed automatically and they will appear under
283 /sys/firmware/acpi/tables.
284 This option turns off this feature.
285 Note that specifying this option does not affect
286 dynamic table installation which will install SSDT
287 tables to /sys/firmware/acpi/tables/dynamic.
289 acpi_rsdp= [ACPI,EFI,KEXEC]
290 Pass the RSDP address to the kernel, mostly used
291 on machines running EFI runtime service to boot the
292 second kernel for kdump.
294 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
295 Format: To spoof as Windows 98: ="Microsoft Windows"
297 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
298 of 2 which is mandated by ACPI 6) as the supported ACPI
299 specification revision (when using this switch, it may
300 be necessary to carry out a cold reboot _twice_ in a
301 row to make it take effect on the platform firmware).
303 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
304 acpi_osi="string1" # add string1
305 acpi_osi="!string2" # remove string2
306 acpi_osi=!* # remove all strings
307 acpi_osi=! # disable all built-in OS vendor
309 acpi_osi= # disable all strings
311 'acpi_osi=!' can be used in combination with single or
312 multiple 'acpi_osi="string1"' to support specific OS
313 vendor string(s). Note that such command can only
314 affect the default state of the OS vendor strings, thus
315 it cannot affect the default state of the feature group
316 strings and the current state of the OS vendor strings,
317 specifying it multiple times through kernel command line
318 is meaningless. This command is useful when one do not
319 care about the state of the feature group strings which
320 should be controlled by the OSPM.
322 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
323 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
324 can make '_OSI("Windows 2000")' TRUE.
326 'acpi_osi=' cannot be used in combination with other
327 'acpi_osi=' command lines, the _OSI method will not
328 exist in the ACPI namespace. NOTE that such command can
329 only affect the _OSI support state, thus specifying it
330 multiple times through kernel command line is also
333 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
336 'acpi_osi=!*' can be used in combination with single or
337 multiple 'acpi_osi="string1"' to support specific
338 string(s). Note that such command can affect the
339 current state of both the OS vendor strings and the
340 feature group strings, thus specifying it multiple times
341 through kernel command line is meaningful. But it may
342 still not able to affect the final state of a string if
343 there are quirks related to this string. This command
344 is useful when one want to control the state of the
345 feature group strings to debug BIOS issues related to
348 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
349 '_OSI("Module Device")' FALSE.
350 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
351 '_OSI("Module Device")' TRUE.
352 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
354 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
356 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
357 they all will make '_OSI("Windows 2000")' TRUE.
360 Override the pmtimer bug detection: force the kernel
361 to assume that this machine's pmtimer latches its value
362 and always returns good values.
364 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
365 Format: { level | edge | high | low }
367 acpi_skip_timer_override [HW,ACPI]
368 Recognize and ignore IRQ0/pin2 Interrupt Override.
369 For broken nForce2 BIOS resulting in XT-PIC timer.
371 acpi_sleep= [HW,ACPI] Sleep options
372 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
373 old_ordering, nonvs, sci_force_enable }
374 See Documentation/power/video.txt for information on
376 s3_beep is for debugging; it makes the PC's speaker beep
377 as soon as the kernel's real-mode entry point is called.
378 s4_nohwsig prevents ACPI hardware signature from being
379 used during resume from hibernation.
380 old_ordering causes the ACPI 1.0 ordering of the _PTS
381 control method, with respect to putting devices into
382 low power states, to be enforced (the ACPI 2.0 ordering
383 of _PTS is used by default).
384 nonvs prevents the kernel from saving/restoring the
385 ACPI NVS memory during suspend/hibernation and resume.
386 sci_force_enable causes the kernel to set SCI_EN directly
387 on resume from S1/S3 (which is against the ACPI spec,
388 but some broken systems don't work without it).
390 acpi_use_timer_override [HW,ACPI]
391 Use timer override. For some broken Nvidia NF5 boards
392 that require a timer override, but don't have HPET
394 add_efi_memmap [EFI; X86] Include EFI memory map in
395 kernel's map of available physical RAM.
398 { off | try_unsupported }
399 off: disable AGP support
400 try_unsupported: try to drive unsupported chipsets
401 (may crash computer or cause data corruption)
404 See Documentation/sound/alsa/alsa-parameters.txt
407 Allow the default userspace alignment fault handler
408 behaviour to be specified. Bit 0 enables warnings,
409 bit 1 enables fixups, and bit 2 sends a segfault.
411 align_va_addr= [X86-64]
412 Align virtual addresses by clearing slice [14:12] when
413 allocating a VMA at process creation time. This option
414 gives you up to 3% performance improvement on AMD F15h
415 machines (where it is enabled by default) for a
416 CPU-intensive style benchmark, and it can vary highly in
417 a microbenchmark depending on workload and compiler.
419 32: only for 32-bit processes
420 64: only for 64-bit processes
421 on: enable for both 32- and 64-bit processes
422 off: disable for both 32- and 64-bit processes
424 alloc_snapshot [FTRACE]
425 Allocate the ftrace snapshot buffer on boot up when the
426 main buffer is allocated. This is handy if debugging
427 and you need to use tracing_snapshot() on boot up, and
428 do not want to use tracing_snapshot_alloc() as it needs
429 to be done where GFP_KERNEL allocations are allowed.
431 amd_iommu= [HW,X86-64]
432 Pass parameters to the AMD IOMMU driver in the system.
434 fullflush - enable flushing of IO/TLB entries when
435 they are unmapped. Otherwise they are
436 flushed before they will be reused, which
438 off - do not initialize any AMD IOMMU found in
440 force_isolation - Force device isolation for all
441 devices. The IOMMU driver is not
442 allowed anymore to lift isolation
443 requirements as needed. This option
444 does not override iommu=pt
446 amd_iommu_dump= [HW,X86-64]
447 Enable AMD IOMMU driver option to dump the ACPI table
448 for AMD IOMMU. With this option enabled, AMD IOMMU
449 driver will print ACPI tables for AMD IOMMU during
450 IOMMU initialization.
452 amijoy.map= [HW,JOY] Amiga joystick support
453 Map of devices attached to JOY0DAT and JOY1DAT
455 See also Documentation/input/joystick.txt
457 analog.map= [HW,JOY] Analog joystick and gamepad support
458 Specifies type or capabilities of an analog joystick
459 connected to one of 16 gameports
460 Format: <type1>,<type2>,..<type16>
463 Power management functions (SPARCstation-4/5 + deriv.)
465 Disable APC CPU standby support. SPARCstation-Fox does
466 not play well with APC CPU idle - disable it if you have
467 APC and your system crashes randomly.
469 apic= [APIC,X86-32] Advanced Programmable Interrupt Controller
470 Change the output verbosity whilst booting
471 Format: { quiet (default) | verbose | debug }
472 Change the amount of debugging information output
473 when initialising the APIC and IO-APIC components.
476 See Documentation/networking/ipv6.txt.
478 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
479 Limit apic dumping. The parameter defines the maximal
480 number of local apics being dumped. Also it is possible
481 to set it to "all" by meaning -- no limit here.
482 Format: { 1 (default) | 2 | ... | all }.
483 The parameter valid if only apic=debug or
484 apic=verbose is specified.
485 Example: apic=debug show_lapic=all
487 apm= [APM] Advanced Power Management
488 See header of arch/x86/kernel/apm_32.c.
490 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
491 Format: <io>,<irq>,<nodeID>
495 atarimouse= [HW,MOUSE] Atari Mouse
497 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
498 EzKey and similar keyboards
500 atkbd.reset= [HW] Reset keyboard during initialization
502 atkbd.set= [HW] Select keyboard code set
503 Format: <int> (2 = AT (default), 3 = PS/2)
505 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
508 atkbd.softraw= [HW] Choose between synthetic and real raw mode
509 Format: <bool> (0 = real, 1 = synthetic (default))
511 atkbd.softrepeat= [HW]
512 Use software keyboard repeat
514 audit= [KNL] Enable the audit sub-system
515 Format: { "0" | "1" } (0 = disabled, 1 = enabled)
516 0 - kernel audit is disabled and can not be enabled
517 until the next reboot
518 unset - kernel audit is initialized but disabled and
519 will be fully enabled by the userspace auditd.
520 1 - kernel audit is initialized and partially enabled,
521 storing at most audit_backlog_limit messages in
522 RAM until it is fully enabled by the userspace
526 audit_backlog_limit= [KNL] Set the audit queue size limit.
527 Format: <int> (must be >=0)
530 baycom_epp= [HW,AX25]
533 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
535 See header of drivers/net/hamradio/baycom_par.c.
537 baycom_ser_fdx= [HW,AX25]
538 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
539 Format: <io>,<irq>,<mode>[,<baud>]
540 See header of drivers/net/hamradio/baycom_ser_fdx.c.
542 baycom_ser_hdx= [HW,AX25]
543 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
544 Format: <io>,<irq>,<mode>
545 See header of drivers/net/hamradio/baycom_ser_hdx.c.
547 blkdevparts= Manual partition parsing of block device(s) for
548 embedded devices based on command line input.
549 See Documentation/block/cmdline-partition.txt
551 boot_delay= Milliseconds to delay each printk during boot.
552 Values larger than 10 seconds (10000) are changed to
556 bootmem_debug [KNL] Enable bootmem allocator debug messages.
558 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
559 bttv.radio= Most important insmod options are available as
561 bttv.pll= See Documentation/video4linux/bttv/Insmod-options
564 bulk_remove=off [PPC] This parameter disables the use of the pSeries
565 firmware feature for flushing multiple hpte entries
568 c101= [NET] Moxa C101 synchronous serial card
570 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
571 Sometimes CPU hardware bugs make them report the cache
572 size incorrectly. The kernel will attempt work arounds
573 to fix known problems, but for some CPUs it is not
574 possible to determine what the correct size should be.
575 This option provides an override for these situations.
577 ca_keys= [KEYS] This parameter identifies a specific key(s) on
578 the system trusted keyring to be used for certificate
580 format: { id:<keyid> | builtin }
582 cca= [MIPS] Override the kernel pages' cache coherency
583 algorithm. Accepted values range from 0 to 7
584 inclusive. See arch/mips/include/asm/pgtable-bits.h
585 for platform specific values (SB1, Loongson3 and
588 ccw_timeout_log [S390]
589 See Documentation/s390/CommonIO for details.
591 cgroup_disable= [KNL] Disable a particular controller
592 Format: {name of the controller(s) to disable}
593 The effects of cgroup_disable=foo are:
594 - foo isn't auto-mounted if you mount all cgroups in
596 - foo isn't visible as an individually mountable
598 {Currently only "memory" controller deal with this and
599 cut the overhead, others just disable the usage. So
600 only cgroup_disable=memory is actually worthy}
602 checkreqprot [SELINUX] Set initial checkreqprot flag value.
603 Format: { "0" | "1" }
604 See security/selinux/Kconfig help text.
605 0 -- check protection applied by kernel (includes
606 any implied execute protection).
607 1 -- check protection requested by application.
608 Default value is set via a kernel config option.
609 Value can be changed at runtime via
610 /selinux/checkreqprot.
613 See Documentation/s390/CommonIO for details.
616 Prevents the clock framework from automatically gating
617 clocks that have not been explicitly enabled by a Linux
618 device driver but are enabled in hardware at reset or
619 by the bootloader/firmware. Note that this does not
620 force such clocks to be always-on nor does it reserve
621 those clocks in any way. This parameter is useful for
622 debug and development, but should not be needed on a
623 platform with proper driver support. For more
624 information, see Documentation/clk.txt.
626 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
628 Forces specified clocksource (if available) to be used
629 when calculating gettimeofday(). If specified
630 clocksource is not available, it defaults to PIT.
631 Format: { pit | tsc | cyclone | pmtmr }
633 clocksource= Override the default clocksource
635 Override the default clocksource and use the clocksource
636 with the name specified.
637 Some clocksource names to choose from, depending on
639 [all] jiffies (this is the base, fallback clocksource)
641 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
642 pxa_timer,timer3,32k_counter,timer0_1
644 [X86-32] pit,hpet,tsc;
645 scx200_hrt on Geode; cyclone on IBM x440
653 clearcpuid=BITNUM [X86]
654 Disable CPUID feature X for the kernel. See
655 arch/x86/include/asm/cpufeatures.h for the valid bit
656 numbers. Note the Linux specific bits are not necessarily
657 stable over kernel options, but the vendor specific
659 Also note that user programs calling CPUID directly
660 or using the feature without checking anything
661 will still see it. This just prevents it from
662 being used by the kernel or shown in /proc/cpuinfo.
663 Also note the kernel might malfunction if you disable
666 cma=nn[MG]@[start[MG][-end[MG]]]
668 Sets the size of kernel global memory area for
669 contiguous memory allocations and optionally the
670 placement constraint by the physical address range of
671 memory allocations. A value of 0 disables CMA
672 altogether. For more information, see
673 include/linux/dma-contiguous.h
675 cmo_free_hint= [PPC] Format: { yes | no }
676 Specify whether pages are marked as being inactive
677 when they are freed. This is used in CMO environments
678 to determine OS memory pressure for page stealing by
682 coherent_pool=nn[KMG] [ARM,KNL]
683 Sets the size of memory pool for coherent, atomic dma
684 allocations, by default set to 256K.
686 code_bytes [X86] How many bytes of object code to print
691 com20020= [HW,NET] ARCnet - COM20020 chipset
693 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
695 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
699 ARCnet - COM90xx chipset (memory-mapped buffers)
700 Format: <io>[,<irq>[,<memstart>]]
702 condev= [HW,S390] console device
705 console= [KNL] Output console device and options.
707 tty<n> Use the virtual console device <n>.
711 Use the specified serial port. The options are of
712 the form "bbbbpnf", where "bbbb" is the baud rate,
713 "p" is parity ("n", "o", or "e"), "n" is number of
714 bits, and "f" is flow control ("r" for RTS or
715 omit it). Default is "9600n8".
717 See Documentation/serial-console.txt for more
719 Documentation/networking/netconsole.txt for an
722 uart[8250],io,<addr>[,options]
723 uart[8250],mmio,<addr>[,options]
724 uart[8250],mmio32,<addr>[,options]
725 uart[8250],0x<addr>[,options]
726 Start an early, polled-mode console on the 8250/16550
727 UART at the specified I/O port or MMIO address,
728 switching to the matching ttyS device later.
729 MMIO inter-register address stride is either 8-bit
730 (mmio) or 32-bit (mmio32).
731 If none of [io|mmio|mmio32], <addr> is assumed to be
732 equivalent to 'mmio'. 'options' are specified in the
733 same format described for ttyS above; if unspecified,
734 the h/w is not re-initialized.
736 hvc<n> Use the hypervisor console device <n>. This is for
737 both Xen and PowerPC hypervisors.
739 If the device connected to the port is not a TTY but a braille
740 device, prepend "brl," before the device type, for instance
742 For now, only VisioBraille is supported.
744 consoleblank= [KNL] The console blank (screen saver) timeout in
745 seconds. Defaults to 10*60 = 10mins. A value of 0
746 disables the blank timer.
749 [KNL] Change the default value for
750 /proc/<pid>/coredump_filter.
751 See also Documentation/filesystems/proc.txt.
753 cpuidle.off=1 [CPU_IDLE]
754 disable the cpuidle sub-system
757 [X86] Delay for N microsec between assert and de-assert
758 of APIC INIT to start processors. This delay occurs
759 on every CPU online, such as boot, and resume from suspend.
762 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
764 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
766 crashkernel=size[KMG][@offset[KMG]]
767 [KNL] Using kexec, Linux can switch to a 'crash kernel'
768 upon panic. This parameter reserves the physical
769 memory region [offset, offset + size] for that kernel
770 image. If '@offset' is omitted, then a suitable offset
771 is selected automatically. Check
772 Documentation/kdump/kdump.txt for further details.
774 crashkernel=range1:size1[,range2:size2,...][@offset]
775 [KNL] Same as above, but depends on the memory
776 in the running system. The syntax of range is
777 start-[end] where start and end are both
778 a memory unit (amount[KMG]). See also
779 Documentation/kdump/kdump.txt for an example.
781 crashkernel=size[KMG],high
782 [KNL, x86_64] range could be above 4G. Allow kernel
783 to allocate physical memory region from top, so could
784 be above 4G if system have more than 4G ram installed.
785 Otherwise memory region will be allocated below 4G, if
787 It will be ignored if crashkernel=X is specified.
788 crashkernel=size[KMG],low
789 [KNL, x86_64] range under 4G. When crashkernel=X,high
790 is passed, kernel could allocate physical memory region
791 above 4G, that cause second kernel crash on system
792 that require some amount of low memory, e.g. swiotlb
793 requires at least 64M+32K low memory, also enough extra
794 low memory is needed to make sure DMA buffers for 32-bit
795 devices won't run out. Kernel would try to allocate at
796 at least 256M below 4G automatically.
797 This one let user to specify own low range under 4G
798 for second kernel instead.
799 0: to disable low allocation.
800 It will be ignored when crashkernel=X,high is not used
801 or memory reserved is below 4G.
806 cs89x0_media= [HW,NET]
807 Format: { rj45 | aui | bnc }
810 See header of drivers/s390/block/dasd_devmap.c.
812 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
813 (one device per port)
814 Format: <port#>,<type>
815 See also Documentation/input/joystick-parport.txt
817 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
818 time. See Documentation/dynamic-debug-howto.txt for
819 details. Deprecated, see dyndbg.
821 debug [KNL] Enable kernel debugging (events log level).
824 [KNL] verbose self-tests
826 Print debugging info while doing the locking API
828 We default to 0 (no extra messages), setting it to
829 1 will print _a lot_ more information - normally
830 only useful to kernel developers.
832 debug_objects [KNL] Enable object debugging
835 [KNL] Disable object debugging
837 debug_guardpage_minorder=
838 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
839 parameter allows control of the order of pages that will
840 be intentionally kept free (and hence protected) by the
841 buddy allocator. Bigger value increase the probability
842 of catching random memory corruption, but reduce the
843 amount of memory for normal system use. The maximum
844 possible value is MAX_ORDER/2. Setting this parameter
845 to 1 or 2 should be enough to identify most random
846 memory corruption problems caused by bugs in kernel or
847 driver code when a CPU writes to (or reads from) a
848 random memory location. Note that there exists a class
849 of memory corruptions problems caused by buggy H/W or
850 F/W or by drivers badly programing DMA (basically when
851 memory is written at bus level and the CPU MMU is
852 bypassed) which are not detectable by
853 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
854 tracking down these problems.
857 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
858 parameter enables the feature at boot time. In
859 default, it is disabled. We can avoid allocating huge
860 chunk of memory for debug pagealloc if we don't enable
861 it at boot time and the system will work mostly same
862 with the kernel built without CONFIG_DEBUG_PAGEALLOC.
863 on: enable the feature
865 debugpat [X86] Enable PAT debugging
867 decnet.addr= [HW,NET]
868 Format: <area>[,<node>]
869 See also Documentation/networking/decnet.txt.
872 [same as hugepagesz=] The size of the default
873 HugeTLB page size. This is the size represented by
874 the legacy /proc/ hugepages APIs, used for SHM, and
875 default size when mounting hugetlbfs filesystems.
876 Defaults to the default architecture's huge page size
880 Set number of hash buckets for dentry cache.
883 See Documentation/networking/ipv6.txt.
885 disable_cpu_apicid= [X86,APIC,SMP]
887 The number of initial APIC ID for the
888 corresponding CPU to be disabled at boot,
889 mostly used for the kdump 2nd kernel to
890 disable BSP to wake up multiple CPUs without
891 causing system reset or hang due to sending
894 disable_ddw [PPC/PSERIES]
895 Disable Dynamic DMA Window support. Use this if
896 to workaround buggy firmware.
899 See Documentation/networking/ipv6.txt.
901 disable_mtrr_cleanup [X86]
902 The kernel tries to adjust MTRR layout from continuous
903 to discrete, to make X server driver able to add WB
904 entry later. This parameter disables that.
906 disable_mtrr_trim [X86, Intel and AMD only]
907 By default the kernel will trim any uncacheable
908 memory out of your available memory pool based on
909 MTRR settings. This parameter disables that behavior,
910 possibly causing your machine to run very slowly.
912 disable_timer_pin_1 [X86]
913 Disable PIN 1 of APIC timer
914 Can be useful to work around chipset bugs.
916 dis_ucode_ldr [X86] Disable the microcode loader.
918 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
919 this option disables the debugging code at boot.
921 dma_debug_entries=<number>
922 This option allows to tune the number of preallocated
923 entries for DMA-API debugging code. One entry is
924 required per DMA-API allocation. Use this if the
925 DMA-API debugging code disables itself because the
926 architectural default is too low.
928 dma_debug_driver=<driver_name>
929 With this option the DMA-API debugging driver
930 filter feature can be enabled at boot time. Just
931 pass the driver to filter for as the parameter.
932 The filter can be disabled or changed to another
933 driver later using sysfs.
935 drm_kms_helper.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
936 Broken monitors, graphic adapters, KVMs and EDIDless
937 panels may send no or incorrect EDID data sets.
938 This parameter allows to specify an EDID data sets
939 in the /lib/firmware directory that are used instead.
940 Generic built-in EDID data sets are used, if one of
941 edid/1024x768.bin, edid/1280x1024.bin,
942 edid/1680x1050.bin, or edid/1920x1080.bin is given
943 and no file with the same name exists. Details and
944 instructions how to build your own EDID data are
945 available in Documentation/EDID/HOWTO.txt. An EDID
946 data set will only be used for a particular connector,
947 if its name and a colon are prepended to the EDID
948 name. Each connector may use a unique EDID data
949 set by separating the files with a comma. An EDID
950 data set with no connector name will be used for
951 any connectors not explicitly specified.
955 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
956 module.dyndbg[="val"]
957 Enable debug messages at boot time. See
958 Documentation/dynamic-debug-howto.txt for details.
960 nompx [X86] Disables Intel Memory Protection Extensions.
961 See Documentation/x86/intel_mpx.txt for more
962 information about the feature.
965 module.async_probe [KNL]
966 Enable asynchronous probe on this module.
968 early_ioremap_debug [KNL]
969 Enable debug messages in early_ioremap support. This
970 is useful for tracking down temporary early mappings
971 which are not unmapped.
973 earlycon= [KNL] Output early console device and options.
975 When used with no options, the early console is
976 determined by the stdout-path property in device
980 Start an early, polled-mode console on a cadence serial
981 port at the specified address. The cadence serial port
982 must already be setup and configured. Options are not
985 uart[8250],io,<addr>[,options]
986 uart[8250],mmio,<addr>[,options]
987 uart[8250],mmio32,<addr>[,options]
988 uart[8250],mmio32be,<addr>[,options]
989 uart[8250],0x<addr>[,options]
990 Start an early, polled-mode console on the 8250/16550
991 UART at the specified I/O port or MMIO address.
992 MMIO inter-register address stride is either 8-bit
993 (mmio) or 32-bit (mmio32 or mmio32be).
994 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
995 to be equivalent to 'mmio'. 'options' are specified
996 in the same format described for "console=ttyS<n>"; if
997 unspecified, the h/w is not initialized.
1000 Start an early, polled-mode console on a pl011 serial
1001 port at the specified address. The pl011 serial port
1002 must already be setup and configured. Options are not
1006 Start an early, polled-mode console on an msm serial
1007 port at the specified address. The serial port
1008 must already be setup and configured. Options are not
1011 msm_serial_dm,<addr>
1012 Start an early, polled-mode console on an msm serial
1013 dm port at the specified address. The serial port
1014 must already be setup and configured. Options are not
1017 smh Use ARM semihosting calls for early console.
1025 Use early console provided by serial driver available
1026 on Samsung SoCs, requires selecting proper type and
1027 a correct base address of the selected UART port. The
1028 serial port must already be setup and configured.
1029 Options are not yet supported.
1033 Use early console provided by Freescale LP UART driver
1034 found on Freescale Vybrid and QorIQ LS1021A processors.
1035 A valid base address must be provided, and the serial
1036 port must already be setup and configured.
1038 earlyprintk= [X86,SH,BLACKFIN,ARM,M68k]
1042 earlyprintk=serial[,ttySn[,baudrate]]
1043 earlyprintk=serial[,0x...[,baudrate]]
1044 earlyprintk=ttySn[,baudrate]
1045 earlyprintk=dbgp[debugController#]
1046 earlyprintk=pciserial,bus:device.function[,baudrate]
1048 earlyprintk is useful when the kernel crashes before
1049 the normal console is initialized. It is not enabled by
1050 default because it has some cosmetic problems.
1052 Append ",keep" to not disable it when the real console
1055 Only one of vga, efi, serial, or usb debug port can
1058 Currently only ttyS0 and ttyS1 may be specified by
1059 name. Other I/O ports may be explicitly specified
1060 on some architectures (x86 and arm at least) by
1061 replacing ttySn with an I/O port address, like this:
1062 earlyprintk=serial,0x1008,115200
1063 You can find the port for a given device in
1064 /proc/tty/driver/serial:
1065 2: uart:ST16650V2 port:00001008 irq:18 ...
1067 Interaction with the standard serial driver is not
1070 The VGA and EFI output is eventually overwritten by
1073 The xen output can only be used by Xen PV guests.
1075 edac_report= [HW,EDAC] Control how to report EDAC event
1076 Format: {"on" | "off" | "force"}
1077 on: enable EDAC to report H/W event. May be overridden
1078 by other higher priority error reporting module.
1079 off: disable H/W event reporting through EDAC.
1080 force: enforce the use of EDAC to report H/W event.
1083 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1086 This is designed to be used in conjunction with
1087 the boot argument: earlyprintk=vga
1090 Format: {"off" | "on" | "skip[mbr]"}
1093 Format: { "old_map", "nochunk", "noruntime", "debug" }
1094 old_map [X86-64]: switch to the old ioremap-based EFI
1095 runtime services mapping. 32-bit still uses this one by
1097 nochunk: disable reading files in "chunks" in the EFI
1098 boot stub, as chunking can cause problems with some
1099 firmware implementations.
1100 noruntime : disable EFI runtime services support
1101 debug: enable misc debug output
1103 efi_no_storage_paranoia [EFI; X86]
1104 Using this parameter you can use more than 50% of
1105 your efi variable storage. Use this parameter only if
1106 you are really sure that your UEFI does sane gc and
1107 fulfills the spec otherwise your board may brick.
1109 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1110 Add arbitrary attribute to specific memory range by
1111 updating original EFI memory map.
1112 Region of memory which aa attribute is added to is
1114 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1115 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1116 attribute is added to range 0x100000000-0x180000000 and
1117 0x10a0000000-0x1120000000.
1119 Using this parameter you can do debugging of EFI memmap
1120 related feature. For example, you can do debugging of
1121 Address Range Mirroring feature even if your box
1124 eisa_irq_edge= [PARISC,HW]
1125 See header of drivers/parisc/eisa.c.
1128 See comment before function elanfreq_setup() in
1129 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1132 Format: {"cfq" | "deadline" | "noop"}
1133 See Documentation/block/cfq-iosched.txt and
1134 Documentation/block/deadline-iosched.txt for details.
1136 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1137 Specifies physical address of start of kernel core
1138 image elf header and optionally the size. Generally
1139 kexec loader will pass this option to capture kernel.
1140 See Documentation/kdump/kdump.txt for details.
1142 enable_mtrr_cleanup [X86]
1143 The kernel tries to adjust MTRR layout from continuous
1144 to discrete, to make X server driver able to add WB
1145 entry later. This parameter enables that.
1147 enable_timer_pin_1 [X86]
1148 Enable PIN 1 of APIC timer
1149 Can be useful to work around chipset bugs
1150 (in particular on some ATI chipsets).
1151 The kernel tries to set a reasonable default.
1153 enforcing [SELINUX] Set initial enforcing status.
1155 See security/selinux/Kconfig help text.
1156 0 -- permissive (log only, no denials).
1157 1 -- enforcing (deny and log).
1159 Value can be changed at runtime via /selinux/enforce.
1162 Disable Error Record Serialization Table (ERST)
1165 ether= [HW,NET] Ethernet cards parameters
1166 This option is obsoleted by the "netdev=" option, which
1167 has equivalent usage. See its documentation for details.
1171 Permit 'security.evm' to be updated regardless of
1172 current integrity status.
1176 fail_make_request=[KNL]
1177 General fault injection mechanism.
1178 Format: <interval>,<probability>,<space>,<times>
1179 See also Documentation/fault-injection/.
1182 See Documentation/blockdev/floppy.txt.
1184 force_pal_cache_flush
1185 [IA-64] Avoid check_sal_cache_flush which may hang on
1186 buggy SAL_CACHE_FLUSH implementations. Using this
1187 parameter will force ia64_sal_cache_flush to call
1188 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1191 Forcefully enable Physical Address Extension (PAE).
1192 Many Pentium M systems disable PAE but may have a
1193 functionally usable PAE implementation.
1194 Warning: use of this parameter will taint the kernel
1195 and may cause unknown problems.
1198 [FTRACE] will set and start the specified tracer
1199 as early as possible in order to facilitate early
1202 ftrace_dump_on_oops[=orig_cpu]
1203 [FTRACE] will dump the trace buffers on oops.
1204 If no parameter is passed, ftrace will dump
1205 buffers of all CPUs, but if you pass orig_cpu, it will
1206 dump only the buffer of the CPU that triggered the
1209 ftrace_filter=[function-list]
1210 [FTRACE] Limit the functions traced by the function
1211 tracer at boot up. function-list is a comma separated
1212 list of functions. This list can be changed at run
1213 time by the set_ftrace_filter file in the debugfs
1216 ftrace_notrace=[function-list]
1217 [FTRACE] Do not trace the functions specified in
1218 function-list. This list can be changed at run time
1219 by the set_ftrace_notrace file in the debugfs
1222 ftrace_graph_filter=[function-list]
1223 [FTRACE] Limit the top level callers functions traced
1224 by the function graph tracer at boot up.
1225 function-list is a comma separated list of functions
1226 that can be changed at run time by the
1227 set_graph_function file in the debugfs tracing directory.
1229 ftrace_graph_notrace=[function-list]
1230 [FTRACE] Do not trace from the functions specified in
1231 function-list. This list is a comma separated list of
1232 functions that can be changed at run time by the
1233 set_graph_notrace file in the debugfs tracing directory.
1236 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1237 support via parallel port (up to 5 devices per port)
1238 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1239 See also Documentation/input/joystick-parport.txt
1243 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1247 gcov_persist= [GCOV] When non-zero (default), profiling data for
1248 kernel modules is saved and remains accessible via
1249 debugfs, even when the module is unloaded/reloaded.
1250 When zero, profiling data is discarded and associated
1251 debugfs files are removed at module unload time.
1253 goldfish [X86] Enable the goldfish android emulator platform.
1254 Don't use this when you are not running on the
1257 gpt [EFI] Forces disk with valid GPT signature but
1258 invalid Protective MBR to be treated as GPT. If the
1259 primary GPT is corrupted, it enables the backup/alternate
1260 GPT to be used instead.
1262 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1263 the "Enable 0" bit of the configuration register.
1266 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1267 the "Enable 0" bit of the configuration register.
1270 grcan.select= [HW] Select which physical interface to use.
1273 grcan.txsize= [HW] Sets the size of the tx buffer.
1274 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1276 grcan.rxsize= [HW] Sets the size of the rx buffer.
1277 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1280 hardlockup_all_cpu_backtrace=
1281 [KNL] Should the hard-lockup detector generate
1282 backtraces on all cpus.
1285 hashdist= [KNL,NUMA] Large hashes allocated during boot
1286 are distributed across NUMA nodes. Defaults on
1287 for 64-bit NUMA, off otherwise.
1288 Format: 0 | 1 (for off | on)
1290 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1292 hd= [EIDE] (E)IDE hard drive subsystem geometry
1293 Format: <cyl>,<head>,<sect>
1296 Disable Hardware Error Source Table (HEST) support;
1297 corresponding firmware-first mode error processing
1298 logic will be disabled.
1300 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1301 size of <nn>. This works even on boxes that have no
1302 highmem otherwise. This also works to reduce highmem
1303 size on bigger boxes.
1305 highres= [KNL] Enable/disable high resolution timer mode.
1306 Valid parameters: "on", "off"
1310 See Documentation/isdn/README.HiSax.
1314 hpet= [X86-32,HPET] option to control HPET usage
1315 Format: { enable (default) | disable | force |
1317 disable: disable HPET and use PIT instead
1318 force: allow force enabled of undocumented chips (ICH4,
1320 verbose: show contents of HPET registers during setup
1322 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1323 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1325 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1326 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1327 On x86-64 and powerpc, this option can be specified
1328 multiple times interleaved with hugepages= to reserve
1329 huge pages of different sizes. Valid pages sizes on
1330 x86-64 are 2M (when the CPU supports "pse") and 1G
1331 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1333 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1334 terminal devices. Valid values: 0..8
1335 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1336 If specified, z/VM IUCV HVC accepts connections
1337 from listed z/VM user IDs only.
1339 hwthread_map= [METAG] Comma-separated list of Linux cpu id to
1340 hardware thread id mappings.
1341 Format: <cpu>:<hwthread>
1344 Do not unregister boot console at start. This is only
1345 useful for debugging when something happens in the window
1346 between unregistering the boot console and initializing
1349 i2c_bus= [HW] Override the default board specific I2C bus speed
1350 or register an additional I2C bus that is not
1351 registered from board initialization code.
1355 i8042.debug [HW] Toggle i8042 debug mode
1356 i8042.unmask_kbd_data
1357 [HW] Enable printing of interrupt data from the KBD port
1358 (disabled by default, and as a pre-condition
1359 requires that i8042.debug=1 be enabled)
1360 i8042.direct [HW] Put keyboard port into non-translated mode
1361 i8042.dumbkbd [HW] Pretend that controller can only read data from
1362 keyboard and cannot control its state
1363 (Don't attempt to blink the leds)
1364 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1365 i8042.nokbd [HW] Don't check/create keyboard port
1366 i8042.noloop [HW] Disable the AUX Loopback command while probing
1368 i8042.nomux [HW] Don't check presence of an active multiplexing
1370 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1372 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1373 i8042.reset [HW] Reset the controller during init, cleanup and
1374 suspend-to-ram transitions, only during s2r
1375 transitions, or never reset
1376 Format: { 1 | Y | y | 0 | N | n }
1377 1, Y, y: always reset controller
1378 0, N, n: don't ever reset controller
1379 Default: only on s2r transitions on x86; most other
1380 architectures force reset to be always executed
1381 i8042.unlock [HW] Unlock (ignore) the keylock
1382 i8042.kbdreset [HW] Reset device connected to KBD port
1386 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1387 indicates that the driver is running on unsupported
1389 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1390 does not match list of supported models.
1392 [HW] Report power status in /proc/i8k
1393 (disabled by default)
1394 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1397 i915.invert_brightness=
1398 [DRM] Invert the sense of the variable that is used to
1399 set the brightness of the panel backlight. Normally a
1400 brightness value of 0 indicates backlight switched off,
1401 and the maximum of the brightness value sets the backlight
1402 to maximum brightness. If this parameter is set to 0
1403 (default) and the machine requires it, or this parameter
1404 is set to 1, a brightness value of 0 sets the backlight
1405 to maximum brightness, and the maximum of the brightness
1406 value switches the backlight off.
1407 -1 -- never invert brightness
1408 0 -- machine default
1409 1 -- force brightness inversion
1412 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1414 ide-core.nodma= [HW] (E)IDE subsystem
1415 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1416 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1417 .cdrom .chs .ignore_cable are additional options
1418 See Documentation/ide/ide.txt.
1420 ide-generic.probe-mask= [HW] (E)IDE subsystem
1422 Probe mask for legacy ISA IDE ports. Depending on
1423 platform up to 6 ports are supported, enabled by
1424 setting corresponding bits in the mask to 1. The
1425 default value is 0x0, which has a special meaning.
1426 On systems that have PCI, it triggers scanning the
1427 PCI bus for the first and the second port, which
1428 are then probed. On systems without PCI the value
1429 of 0x0 enables probing the two first ports as if it
1432 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1433 Claim all unknown PCI IDE storage controllers.
1436 Format: idle=poll, idle=halt, idle=nomwait
1437 Poll forces a polling idle loop that can slightly
1438 improve the performance of waking up a idle CPU, but
1439 will use a lot of power and make the system run hot.
1441 idle=halt: Halt is forced to be used for CPU idle.
1442 In such case C2/C3 won't be used again.
1443 idle=nomwait: Disable mwait for CPU C-states
1445 ignore_loglevel [KNL]
1446 Ignore loglevel setting - this will print /all/
1447 kernel messages to the console. Useful for debugging.
1448 We also add it as printk module parameter, so users
1449 could change it dynamically, usually by
1450 /sys/module/printk/parameters/ignore_loglevel.
1452 ihash_entries= [KNL]
1453 Set number of hash buckets for inode cache.
1455 ima_appraise= [IMA] appraise integrity measurements
1456 Format: { "off" | "enforce" | "fix" | "log" }
1459 ima_appraise_tcb [IMA]
1460 The builtin appraise policy appraises all files
1464 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1468 The list of supported hash algorithms is defined
1469 in crypto/hash_info.h.
1472 The builtin measurement policy to load during IMA
1473 setup. Specyfing "tcb" as the value, measures all
1474 programs exec'd, files mmap'd for exec, and all files
1475 opened with the read mode bit set by either the
1476 effective uid (euid=0) or uid=0.
1479 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1480 Load a policy which meets the needs of the Trusted
1481 Computing Base. This means IMA will measure all
1482 programs exec'd, files mmap'd for exec, and all files
1483 opened for read by uid=0.
1486 Select one of defined IMA measurements template formats.
1487 Formats: { "ima" | "ima-ng" | "ima-sig" }
1491 [IMA] Define a custom template format.
1492 Format: { "field1|...|fieldN" }
1494 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1495 Format: <min_file_size>
1496 Set the minimal file size for using asynchronous hash.
1497 If left unspecified, ahash usage is disabled.
1499 ahash performance varies for different data sizes on
1500 different crypto accelerators. This option can be used
1501 to achieve the best performance for a particular HW.
1503 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1505 Set hashing buffer size. Default: 4k.
1507 ahash performance varies for different chunk sizes on
1508 different crypto accelerators. This option can be used
1509 to achieve best performance for particular HW.
1513 Run specified binary instead of /sbin/init as init
1516 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1517 for working out where the kernel is dying during
1520 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1521 initcall functions. Useful for debugging built-in
1522 modules and initcalls.
1524 initrd= [BOOT] Specify the location of the initial ramdisk
1526 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1529 int_pln_enable [x86] Enable power limit notification interrupt
1531 integrity_audit=[IMA]
1532 Format: { "0" | "1" }
1533 0 -- basic integrity auditing messages. (Default)
1534 1 -- additional integrity auditing messages.
1536 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1538 Enable intel iommu driver.
1540 Disable intel iommu driver.
1541 igfx_off [Default Off]
1542 By default, gfx is mapped as normal device. If a gfx
1543 device has a dedicated DMAR unit, the DMAR unit is
1544 bypassed by not enabling DMAR with this option. In
1545 this case, gfx device will use physical address for
1548 With this option iommu will not optimize to look
1549 for io virtual address below 32-bit forcing dual
1550 address cycle on pci bus for cards supporting greater
1551 than 32-bit addressing. The default is to look
1552 for translation below 32-bit and if not available
1553 then look in the higher range.
1554 strict [Default Off]
1555 With this option on every unmap_single operation will
1556 result in a hardware IOTLB flush operation as opposed
1557 to batching them for performance.
1558 sp_off [Default Off]
1559 By default, super page will be supported if Intel IOMMU
1560 has the capability. With this option, super page will
1562 ecs_off [Default Off]
1563 By default, extended context tables will be supported if
1564 the hardware advertises that it has support both for the
1565 extended tables themselves, and also PASID support. With
1566 this option set, extended tables will not be used even
1567 on hardware which claims to support them.
1569 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1570 0 disables intel_idle and fall back on acpi_idle.
1571 1 to 6 specify maximum depth of C-state.
1575 Do not enable intel_pstate as the default
1576 scaling driver for the supported processors
1578 Enable intel_pstate on systems that prohibit it by default
1579 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1580 instead of acpi-cpufreq may disable platform features, such
1581 as thermal controls and power capping, that rely on ACPI
1582 P-States information being indicated to OSPM and therefore
1583 should be used with caution. This option does not work with
1584 processors that aren't supported by the intel_pstate driver
1585 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1587 Do not enable hardware P state control (HWP)
1590 Only load intel_pstate on systems which support
1591 hardware P state control (HWP) if available.
1593 intremap= [X86-64, Intel-IOMMU]
1594 on enable Interrupt Remapping (default)
1595 off disable Interrupt Remapping
1596 nosid disable Source ID checking
1598 BIOS x2APIC opt-out request will be ignored
1599 nopost disable Interrupt Posting
1601 iomem= Disable strict checking of access to MMIO memory
1602 strict regions from userspace.
1617 nobypass [PPC/POWERNV]
1618 Disable IOMMU bypass, using IOMMU for PCI devices.
1621 io7= [HW] IO7 for Marvel based alpha systems
1622 See comment before marvel_specify_io7 in
1623 arch/alpha/kernel/core_marvel.c.
1625 io_delay= [X86] I/O delay method
1627 Standard port 0x80 based delay
1629 Alternate port 0xed based delay (needed on some systems)
1631 Simple two microseconds delay
1636 See Documentation/filesystems/nfs/nfsroot.txt.
1639 When an interrupt is not handled search all handlers
1640 for it. Intended to get systems with badly broken
1644 When an interrupt is not handled search all handlers
1645 for it. Also check all handlers each timer
1646 interrupt. Intended to get systems with badly broken
1650 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1652 isolcpus= [KNL,SMP] Isolate CPUs from the general scheduler.
1654 <cpu number>,...,<cpu number>
1656 <cpu number>-<cpu number>
1657 (must be a positive range in ascending order)
1659 <cpu number>,...,<cpu number>-<cpu number>
1661 This option can be used to specify one or more CPUs
1662 to isolate from the general SMP balancing and scheduling
1663 algorithms. You can move a process onto or off an
1664 "isolated" CPU via the CPU affinity syscalls or cpuset.
1665 <cpu number> begins at 0 and the maximum value is
1666 "number of CPUs in system - 1".
1668 This option is the preferred way to isolate CPUs. The
1669 alternative -- manually setting the CPU mask of all
1670 tasks in the system -- can cause problems and
1671 suboptimal load balancer performance.
1675 ivrs_ioapic [HW,X86_64]
1676 Provide an override to the IOAPIC-ID<->DEVICE-ID
1677 mapping provided in the IVRS ACPI table. For
1678 example, to map IOAPIC-ID decimal 10 to
1679 PCI device 00:14.0 write the parameter as:
1680 ivrs_ioapic[10]=00:14.0
1682 ivrs_hpet [HW,X86_64]
1683 Provide an override to the HPET-ID<->DEVICE-ID
1684 mapping provided in the IVRS ACPI table. For
1685 example, to map HPET-ID decimal 0 to
1686 PCI device 00:14.0 write the parameter as:
1687 ivrs_hpet[0]=00:14.0
1689 js= [HW,JOY] Analog joystick
1690 See Documentation/input/joystick.txt.
1693 Enable/disable kernel and module base offset ASLR
1694 (Address Space Layout Randomization) if built into
1695 the kernel. When CONFIG_HIBERNATION is selected,
1696 kASLR is disabled by default. When kASLR is enabled,
1697 hibernation will be disabled.
1701 kernelcore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
1702 specifies the amount of memory usable by the kernel
1703 for non-movable allocations. The requested amount is
1704 spread evenly throughout all nodes in the system. The
1705 remaining memory in each node is used for Movable
1706 pages. In the event, a node is too small to have both
1707 kernelcore and Movable pages, kernelcore pages will
1708 take priority and other nodes will have a larger number
1709 of Movable pages. The Movable zone is used for the
1710 allocation of pages that may be reclaimed or moved
1711 by the page migration subsystem. This means that
1712 HugeTLB pages may not be allocated from this zone.
1713 Note that allocations like PTEs-from-HighMem still
1714 use the HighMem zone if it exists, and the Normal
1715 zone if it does not.
1717 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1718 Format: <Controller#>[,poll interval]
1719 The controller # is the number of the ehci usb debug
1720 port as it is probed via PCI. The poll interval is
1721 optional and is the number seconds in between
1722 each poll cycle to the debug port in case you need
1723 the functionality for interrupting the kernel with
1724 gdb or control-c on the dbgp connection. When
1725 not using this parameter you use sysrq-g to break into
1726 the kernel debugger.
1728 kgdboc= [KGDB,HW] kgdb over consoles.
1729 Requires a tty driver that supports console polling,
1730 or a supported polling keyboard driver (non-usb).
1731 Serial only format: <serial_device>[,baud]
1732 keyboard only format: kbd
1733 keyboard and serial format: kbd,<serial_device>[,baud]
1734 Optional Kernel mode setting:
1735 kms, kbd format: kms,kbd
1736 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1738 kgdbwait [KGDB] Stop kernel execution and enter the
1739 kernel debugger at the earliest opportunity.
1741 kmac= [MIPS] korina ethernet MAC address.
1742 Configure the RouterBoard 532 series on-chip
1743 Ethernet adapter MAC address.
1745 kmemleak= [KNL] Boot-time kmemleak enable/disable
1746 Valid arguments: on, off
1748 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
1751 kmemcheck= [X86] Boot-time kmemcheck enable/disable/one-shot mode
1752 Valid arguments: 0, 1, 2
1753 kmemcheck=0 (disabled)
1754 kmemcheck=1 (enabled)
1755 kmemcheck=2 (one-shot mode)
1756 Default: 2 (one-shot mode)
1758 kstack=N [X86] Print N words from the kernel stack
1761 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1762 Default is 0 (don't ignore, but inject #GP)
1764 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
1768 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
1769 Default is 1 (enabled)
1771 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
1773 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
1775 kvm-intel.ept= [KVM,Intel] Disable extended page tables
1776 (virtualized MMU) support on capable Intel chips.
1777 Default is 1 (enabled)
1779 kvm-intel.emulate_invalid_guest_state=
1780 [KVM,Intel] Enable emulation of invalid guest states
1781 Default is 0 (disabled)
1783 kvm-intel.flexpriority=
1784 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
1785 Default is 1 (enabled)
1788 [KVM,Intel] Enable VMX nesting (nVMX).
1789 Default is 0 (disabled)
1791 kvm-intel.unrestricted_guest=
1792 [KVM,Intel] Disable unrestricted guest feature
1793 (virtualized real and unpaged mode) on capable
1794 Intel chips. Default is 1 (enabled)
1796 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
1797 feature (tagged TLBs) on capable Intel chips.
1798 Default is 1 (enabled)
1804 lapic [X86-32,APIC] Enable the local APIC even if BIOS
1807 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
1808 value for LAPIC timer one-shot implementation. Default
1809 back to the programmable timer unit in the LAPIC.
1811 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
1814 libata.dma= [LIBATA] DMA control
1815 libata.dma=0 Disable all PATA and SATA DMA
1816 libata.dma=1 PATA and SATA Disk DMA only
1817 libata.dma=2 ATAPI (CDROM) DMA only
1818 libata.dma=4 Compact Flash DMA only
1819 Combinations also work, so libata.dma=3 enables DMA
1820 for disks and CDROMs, but not CFs.
1822 libata.ignore_hpa= [LIBATA] Ignore HPA limit
1823 libata.ignore_hpa=0 keep BIOS limits (default)
1824 libata.ignore_hpa=1 ignore limits, using full disk
1826 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
1830 libata.force= [LIBATA] Force configurations. The format is comma
1831 separated list of "[ID:]VAL" where ID is
1832 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
1833 matching port, link or device. Basically, it matches
1834 the ATA ID string printed on console by libata. If
1835 the whole ID part is omitted, the last PORT and DEVICE
1836 values are used. If ID hasn't been specified yet, the
1837 configuration applies to all ports, links and devices.
1839 If only DEVICE is omitted, the parameter applies to
1840 the port and all links and devices behind it. DEVICE
1841 number of 0 either selects the first device or the
1842 first fan-out link behind PMP device. It does not
1843 select the host link. DEVICE number of 15 selects the
1844 host link and device attached to it.
1846 The VAL specifies the configuration to force. As long
1847 as there's no ambiguity shortcut notation is allowed.
1848 For example, both 1.5 and 1.5G would work for 1.5Gbps.
1849 The following configurations can be forced.
1851 * Cable type: 40c, 80c, short40c, unk, ign or sata.
1852 Any ID with matching PORT is used.
1854 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
1856 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
1857 udma[/][16,25,33,44,66,100,133] notation is also
1860 * [no]ncq: Turn on or off NCQ.
1862 * [no]ncqtrim: Turn off queued DSM TRIM.
1864 * nohrst, nosrst, norst: suppress hard, soft
1867 * rstonce: only attempt one reset during
1868 hot-unplug link recovery
1870 * dump_id: dump IDENTIFY data.
1872 * atapi_dmadir: Enable ATAPI DMADIR bridge support
1874 * disable: Disable this device.
1876 If there are multiple matching configurations changing
1877 the same attribute, the last one is used.
1879 memblock=debug [KNL] Enable memblock debug messages.
1881 load_ramdisk= [RAM] List of ramdisks to load from floppy
1882 See Documentation/blockdev/ramdisk.txt.
1884 lockd.nlm_grace_period=P [NFS] Assign grace period.
1887 lockd.nlm_tcpport=N [NFS] Assign TCP port.
1890 lockd.nlm_timeout=T [NFS] Assign timeout value.
1893 lockd.nlm_udpport=M [NFS] Assign UDP port.
1896 locktorture.nreaders_stress= [KNL]
1897 Set the number of locking read-acquisition kthreads.
1898 Defaults to being automatically set based on the
1899 number of online CPUs.
1901 locktorture.nwriters_stress= [KNL]
1902 Set the number of locking write-acquisition kthreads.
1904 locktorture.onoff_holdoff= [KNL]
1905 Set time (s) after boot for CPU-hotplug testing.
1907 locktorture.onoff_interval= [KNL]
1908 Set time (s) between CPU-hotplug operations, or
1909 zero to disable CPU-hotplug testing.
1911 locktorture.shuffle_interval= [KNL]
1912 Set task-shuffle interval (jiffies). Shuffling
1913 tasks allows some CPUs to go into dyntick-idle
1914 mode during the locktorture test.
1916 locktorture.shutdown_secs= [KNL]
1917 Set time (s) after boot system shutdown. This
1918 is useful for hands-off automated testing.
1920 locktorture.stat_interval= [KNL]
1921 Time (s) between statistics printk()s.
1923 locktorture.stutter= [KNL]
1924 Time (s) to stutter testing, for example,
1925 specifying five seconds causes the test to run for
1926 five seconds, wait for five seconds, and so on.
1927 This tests the locking primitive's ability to
1928 transition abruptly to and from idle.
1930 locktorture.torture_runnable= [BOOT]
1931 Start locktorture running at boot time.
1933 locktorture.torture_type= [KNL]
1934 Specify the locking implementation to test.
1936 locktorture.verbose= [KNL]
1937 Enable additional printk() statements.
1939 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
1942 loglevel= All Kernel Messages with a loglevel smaller than the
1943 console loglevel will be printed to the console. It can
1944 also be changed with klogd or other programs. The
1945 loglevels are defined as follows:
1947 0 (KERN_EMERG) system is unusable
1948 1 (KERN_ALERT) action must be taken immediately
1949 2 (KERN_CRIT) critical conditions
1950 3 (KERN_ERR) error conditions
1951 4 (KERN_WARNING) warning conditions
1952 5 (KERN_NOTICE) normal but significant condition
1953 6 (KERN_INFO) informational
1954 7 (KERN_DEBUG) debug-level messages
1956 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
1957 in bytes. n must be a power of two and greater
1958 than the minimal size. The minimal size is defined
1959 by LOG_BUF_SHIFT kernel config parameter. There is
1960 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
1961 that allows to increase the default size depending on
1962 the number of CPUs. See init/Kconfig for more details.
1964 logo.nologo [FB] Disables display of the built-in Linux logo.
1965 This may be used to provide more screen space for
1966 kernel log messages and is useful when debugging
1967 kernel boot problems.
1969 lp=0 [LP] Specify parallel ports to use, e.g,
1970 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
1971 lp=reset first parallel port). 'lp=0' disables the
1972 lp=auto printer driver. 'lp=reset' (which can be
1973 specified in addition to the ports) causes
1974 attached printers to be reset. Using
1975 lp=port1,port2,... specifies the parallel ports
1976 to associate lp devices with, starting with
1977 lp0. A port specification may be 'none' to skip
1978 that lp device, or a parport name such as
1979 'parport0'. Specifying 'lp=auto' instead of a
1980 port specification list means that device IDs
1981 from each port should be examined, to see if
1982 an IEEE 1284-compliant printer is attached; if
1983 so, the driver will manage that printer.
1984 See also header of drivers/char/lp.c.
1987 Sets loops_per_jiffy to given constant, thus avoiding
1988 time-consuming boot-time autodetection (up to 250 ms per
1989 CPU). 0 enables autodetection (default). To determine
1990 the correct value for your kernel, boot with normal
1991 autodetection and see what value is printed. Note that
1992 on SMP systems the preset will be applied to all CPUs,
1993 which is likely to cause problems if your CPUs need
1994 significantly divergent settings. An incorrect value
1995 will cause delays in the kernel to be wrong, leading to
1996 unpredictable I/O errors and other breakage. Although
1997 unlikely, in the extreme case this might damage your
2001 Format: <io>,<irq>,<dma>
2003 machvec= [IA-64] Force the use of a particular machine-vector
2004 (machvec) in a generic kernel.
2005 Example: machvec=hpzx1_swiotlb
2007 machtype= [Loongson] Share the same kernel image file between different
2009 Example: machtype=lemote-yeeloong-2f-7inch
2011 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2012 than or equal to this physical address is ignored.
2014 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2015 should make use of. maxcpus=n : n >= 0 limits the
2016 kernel to using 'n' processors. n=0 is a special case,
2017 it is equivalent to "nosmp", which also disables
2020 max_loop= [LOOP] The number of loop block devices that get
2021 (loop.max_loop) unconditionally pre-created at init time. The default
2022 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2023 of statically allocating a predefined number, loop
2024 devices can be requested on-demand with the
2025 /dev/loop-control interface.
2027 mce [X86-32] Machine Check Exception
2029 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
2031 md= [HW] RAID subsystems devices and level
2032 See Documentation/md.txt.
2035 Format: <first>,<last>
2036 Specifies range of consoles to be captured by the MDA.
2039 Control mitigation for the Micro-architectural Data
2040 Sampling (MDS) vulnerability.
2042 Certain CPUs are vulnerable to an exploit against CPU
2043 internal buffers which can forward information to a
2044 disclosure gadget under certain conditions.
2046 In vulnerable processors, the speculatively
2047 forwarded data can be used in a cache side channel
2048 attack, to access data to which the attacker does
2049 not have direct access.
2051 This parameter controls the MDS mitigation. The
2054 full - Enable MDS mitigation on vulnerable CPUs
2055 off - Unconditionally disable MDS mitigation
2057 On TAA-affected machines, mds=off can be prevented by
2058 an active TAA mitigation as both vulnerabilities are
2059 mitigated with the same mechanism so in order to disable
2060 this mitigation, you need to specify tsx_async_abort=off
2063 Not specifying this option is equivalent to
2066 For details see: Documentation/hw-vuln/mds.rst
2068 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2069 Amount of memory to be used when the kernel is not able
2070 to see the whole system memory or for test.
2071 [X86] Work as limiting max address. Use together
2072 with memmap= to avoid physical address space collisions.
2073 Without memmap= PCI devices could be placed at addresses
2074 belonging to unused RAM.
2076 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2080 [KNL,SH] Allow user to override the default size for
2081 per-device physically contiguous DMA buffers.
2083 memmap=exactmap [KNL,X86] Enable setting of an exact
2084 E820 memory map, as specified by the user.
2085 Such memmap=exactmap lines can be constructed based on
2086 BIOS output or other requirements. See the memmap=nn@ss
2089 memmap=nn[KMG]@ss[KMG]
2090 [KNL] Force usage of a specific region of memory.
2091 Region of memory to be used is from ss to ss+nn.
2093 memmap=nn[KMG]#ss[KMG]
2094 [KNL,ACPI] Mark specific memory as ACPI data.
2095 Region of memory to be marked is from ss to ss+nn.
2097 memmap=nn[KMG]$ss[KMG]
2098 [KNL,ACPI] Mark specific memory as reserved.
2099 Region of memory to be reserved is from ss to ss+nn.
2100 Example: Exclude memory from 0x18690000-0x1869ffff
2101 memmap=64K$0x18690000
2103 memmap=0x10000$0x18690000
2105 memmap=nn[KMG]!ss[KMG]
2106 [KNL,X86] Mark specific memory as protected.
2107 Region of memory to be used, from ss to ss+nn.
2108 The memory region may be marked as e820 type 12 (0xc)
2109 and is NVDIMM or ADR memory.
2111 memory_corruption_check=0/1 [X86]
2112 Some BIOSes seem to corrupt the first 64k of
2113 memory when doing things like suspend/resume.
2114 Setting this option will scan the memory
2115 looking for corruption. Enabling this will
2116 both detect corruption and prevent the kernel
2117 from using the memory being corrupted.
2118 However, its intended as a diagnostic tool; if
2119 repeatable BIOS-originated corruption always
2120 affects the same memory, you can use memmap=
2121 to prevent the kernel from using that memory.
2123 memory_corruption_check_size=size [X86]
2124 By default it checks for corruption in the low
2125 64k, making this memory unavailable for normal
2126 use. Use this parameter to scan for
2127 corruption in more or less memory.
2129 memory_corruption_check_period=seconds [X86]
2130 By default it checks for corruption every 60
2131 seconds. Use this parameter to check at some
2132 other rate. 0 disables periodic checking.
2134 memtest= [KNL,X86,ARM] Enable memtest
2136 default : 0 <disable>
2137 Specifies the number of memtest passes to be
2138 performed. Each pass selects another test
2139 pattern from a given set of patterns. Memtest
2140 fills the memory with this pattern, validates
2141 memory contents and reserves bad memory
2142 regions that are detected.
2144 meye.*= [HW] Set MotionEye Camera parameters
2145 See Documentation/video4linux/meye.txt.
2147 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2148 Multi-Function General Purpose Timers on AMD Geode
2151 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2152 the BIOS has incorrectly applied a workaround. TinyBIOS
2153 version 0.98 is known to be affected, 0.99 fixes the
2154 problem by letting the user disable the workaround.
2158 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2159 physical address is ignored.
2161 mini2440= [ARM,HW,KNL]
2162 Format:[0..2][b][c][t]
2164 MINI2440 configuration specification:
2165 0 - The attached screen is the 3.5" TFT
2166 1 - The attached screen is the 7" TFT
2167 2 - The VGA Shield is attached (1024x768)
2168 Leaving out the screen size parameter will not load
2169 the TFT driver, and the framebuffer will be left
2171 b - Enable backlight. The TFT backlight pin will be
2172 linked to the kernel VESA blanking code and a GPIO
2173 LED. This parameter is not necessary when using the
2175 c - Enable the s3c camera interface.
2176 t - Reserved for enabling touchscreen support. The
2177 touchscreen support is not enabled in the mainstream
2178 kernel as of 2.6.30, a preliminary port can be found
2179 in the "bleeding edge" mini2440 support kernel at
2180 http://repo.or.cz/w/linux-2.6/mini2440.git
2183 [X86] Control optional mitigations for CPU
2184 vulnerabilities. This is a set of curated,
2185 arch-independent options, each of which is an
2186 aggregation of existing arch-specific options.
2189 Disable all optional CPU mitigations. This
2190 improves system performance, but it may also
2191 expose users to several CPU vulnerabilities.
2192 Equivalent to: nopti [X86]
2195 spectre_v2_user=off [X86]
2196 spec_store_bypass_disable=off [X86]
2198 tsx_async_abort=off [X86]
2199 no_entry_flush [PPC]
2200 no_uaccess_flush [PPC]
2203 Mitigate all CPU vulnerabilities, but leave SMT
2204 enabled, even if it's vulnerable. This is for
2205 users who don't want to be surprised by SMT
2206 getting disabled across kernel upgrades, or who
2207 have other ways of avoiding SMT-based attacks.
2208 Equivalent to: (default behavior)
2211 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2212 parameter allows control of the logging verbosity for
2213 the additional memory initialisation checks. A value
2214 of 0 disables mminit logging and a level of 4 will
2215 log everything. Information is printed at KERN_DEBUG
2216 so loglevel=8 may also need to be specified.
2219 [KNL] When CONFIG_MODULE_SIG is set, this means that
2220 modules without (valid) signatures will fail to load.
2221 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2222 is always true, so this option does nothing.
2225 [MOUSE] Maximum time between finger touching and
2226 leaving touchpad surface for touch to be considered
2227 a tap and be reported as a left button click (for
2228 touchpads working in absolute mode only).
2230 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2231 reporting absolute coordinates, such as tablets
2232 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2233 reporting absolute coordinates, such as tablets
2235 movablecore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
2236 is similar to kernelcore except it specifies the
2237 amount of memory used for migratable allocations.
2238 If both kernelcore and movablecore is specified,
2239 then kernelcore will be at *least* the specified
2240 value but may be more. If movablecore on its own
2241 is specified, the administrator must be careful
2242 that the amount of memory usable for all allocations
2245 movable_node [KNL,X86] Boot-time switch to enable the effects
2246 of CONFIG_MOVABLE_NODE=y. See mm/Kconfig for details.
2248 MTD_Partition= [MTD]
2249 Format: <name>,<region-number>,<size>,<offset>
2251 MTD_Region= [MTD] Format:
2252 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2255 See drivers/mtd/cmdlinepart.c.
2257 multitce=off [PPC] This parameter disables the use of the pSeries
2258 firmware feature for updating multiple TCE entries
2261 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2263 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2265 boundary - index of last SLC block on Flex-OneNAND.
2266 The remaining blocks are configured as MLC blocks.
2267 lock - Configure if Flex-OneNAND boundary should be locked.
2268 Once locked, the boundary cannot be changed.
2269 1 indicates lock status, 0 indicates unlock status.
2272 ARM/S3C2412 JIVE boot control
2274 See arch/arm/mach-s3c2412/mach-jive.c
2276 mtouchusb.raw_coordinates=
2277 [HW] Make the MicroTouch USB driver use raw coordinates
2278 ('y', default) or cooked coordinates ('n')
2280 mtrr_chunk_size=nn[KMG] [X86]
2281 used for mtrr cleanup. It is largest continuous chunk
2282 that could hold holes aka. UC entries.
2284 mtrr_gran_size=nn[KMG] [X86]
2285 Used for mtrr cleanup. It is granularity of mtrr block.
2287 Large value could prevent small alignment from
2290 mtrr_spare_reg_nr=n [X86]
2292 Range: 0,7 : spare reg number
2294 Used for mtrr cleanup. It is spare mtrr entries number.
2295 Set to 2 or more if your graphical card needs more.
2297 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2299 netdev= [NET] Network devices parameters
2300 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2301 Note that mem_start is often overloaded to mean
2302 something different and driver-specific.
2303 This usage is only documented in each driver source
2307 [NETFILTER] Enable connection tracking flow accounting
2308 0 to disable accounting
2309 1 to enable accounting
2312 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2313 See Documentation/filesystems/nfs/nfsroot.txt.
2315 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2316 See Documentation/filesystems/nfs/nfsroot.txt.
2318 nfsrootdebug [NFS] enable nfsroot debugging messages.
2319 See Documentation/filesystems/nfs/nfsroot.txt.
2321 nfs.callback_tcpport=
2322 [NFS] set the TCP port on which the NFSv4 callback
2323 channel should listen.
2326 [NFS] sets the pathname to the program which is used
2327 to update the NFS client cache entries.
2329 nfs.cache_getent_timeout=
2330 [NFS] sets the timeout after which an attempt to
2331 update a cache entry is deemed to have failed.
2333 nfs.idmap_cache_timeout=
2334 [NFS] set the maximum lifetime for idmapper cache
2338 [NFS] enable 64-bit inode numbers.
2339 If zero, the NFS client will fake up a 32-bit inode
2340 number for the readdir() and stat() syscalls instead
2341 of returning the full 64-bit number.
2342 The default is to return 64-bit inode numbers.
2344 nfs.max_session_slots=
2345 [NFSv4.1] Sets the maximum number of session slots
2346 the client will attempt to negotiate with the server.
2347 This limits the number of simultaneous RPC requests
2348 that the client can send to the NFSv4.1 server.
2349 Note that there is little point in setting this
2350 value higher than the max_tcp_slot_table_limit.
2352 nfs.nfs4_disable_idmapping=
2353 [NFSv4] When set to the default of '1', this option
2354 ensures that both the RPC level authentication
2355 scheme and the NFS level operations agree to use
2356 numeric uids/gids if the mount is using the
2357 'sec=sys' security flavour. In effect it is
2358 disabling idmapping, which can make migration from
2359 legacy NFSv2/v3 systems to NFSv4 easier.
2360 Servers that do not support this mode of operation
2361 will be autodetected by the client, and it will fall
2362 back to using the idmapper.
2363 To turn off this behaviour, set the value to '0'.
2365 [NFS4] Specify an additional fixed unique ident-
2366 ification string that NFSv4 clients can insert into
2367 their nfs_client_id4 string. This is typically a
2368 UUID that is generated at system install time.
2370 nfs.send_implementation_id =
2371 [NFSv4.1] Send client implementation identification
2372 information in exchange_id requests.
2373 If zero, no implementation identification information
2375 The default is to send the implementation identification
2378 nfs.recover_lost_locks =
2379 [NFSv4] Attempt to recover locks that were lost due
2380 to a lease timeout on the server. Please note that
2381 doing this risks data corruption, since there are
2382 no guarantees that the file will remain unchanged
2383 after the locks are lost.
2384 If you want to enable the kernel legacy behaviour of
2385 attempting to recover these locks, then set this
2387 The default parameter value of '0' causes the kernel
2388 not to attempt recovery of lost locks.
2390 nfs4.layoutstats_timer =
2391 [NFSv4.2] Change the rate at which the kernel sends
2392 layoutstats to the pNFS metadata server.
2394 Setting this to value to 0 causes the kernel to use
2395 whatever value is the default set by the layout
2396 driver. A non-zero value sets the minimum interval
2397 in seconds between layoutstats transmissions.
2399 nfsd.nfs4_disable_idmapping=
2400 [NFSv4] When set to the default of '1', the NFSv4
2401 server will return only numeric uids and gids to
2402 clients using auth_sys, and will accept numeric uids
2403 and gids from such clients. This is intended to ease
2404 migration from NFSv2/v3.
2406 objlayoutdriver.osd_login_prog=
2407 [NFS] [OBJLAYOUT] sets the pathname to the program which
2408 is used to automatically discover and login into new
2409 osd-targets. Please see:
2410 Documentation/filesystems/pnfs.txt for more explanations
2412 nmi_debug= [KNL,AVR32,SH] Specify one or more actions to take
2413 when a NMI is triggered.
2414 Format: [state][,regs][,debounce][,die]
2416 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2417 Format: [panic,][nopanic,][num]
2419 0 - turn hardlockup detector in nmi_watchdog off
2420 1 - turn hardlockup detector in nmi_watchdog on
2421 When panic is specified, panic when an NMI watchdog
2422 timeout occurs (or 'nopanic' to override the opposite
2423 default). To disable both hard and soft lockup detectors,
2424 please see 'nowatchdog'.
2425 This is useful when you use a panic=... timeout and
2426 need the box quickly up again.
2428 netpoll.carrier_timeout=
2429 [NET] Specifies amount of time (in seconds) that
2430 netpoll should wait for a carrier. By default netpoll
2433 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2434 emulation library even if a 387 maths coprocessor
2438 [HW] Never suspend the console
2439 Disable suspending of consoles during suspend and
2440 hibernate operations. Once disabled, debugging
2441 messages can reach various consoles while the rest
2442 of the system is being put to sleep (ie, while
2443 debugging driver suspend/resume hooks). This may
2444 not work reliably with all consoles, but is known
2445 to work with serial and VGA consoles.
2446 To facilitate more flexible debugging, we also add
2447 console_suspend, a printk module parameter to control
2448 it. Users could use console_suspend (usually
2449 /sys/module/printk/parameters/console_suspend) to
2450 turn on/off it dynamically.
2452 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
2453 caches in the slab allocator. Saves per-node memory,
2454 but will impact performance.
2458 noaltinstr [S390] Disables alternative instructions patching
2459 (CPU alternatives feature).
2461 noapic [SMP,APIC] Tells the kernel to not make use of any
2462 IOAPICs that may be present in the system.
2464 noautogroup Disable scheduler automatic task group creation.
2466 nobats [PPC] Do not use BATs for mapping kernel lowmem
2467 on "Classic" PPC cores.
2471 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
2473 nodelayacct [KNL] Disable per-task delay accounting
2475 nodisconnect [HW,SCSI,M68K] Disables SCSI disconnects.
2477 nodsp [SH] Disable hardware DSP at boot time.
2479 noefi Disable EFI runtime services support.
2481 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
2486 On X86-32 available only on PAE configured kernels.
2487 noexec=on: enable non-executable mappings (default)
2488 noexec=off: disable non-executable mappings
2491 Disable SMAP (Supervisor Mode Access Prevention)
2492 even if it is supported by processor.
2495 Disable SMEP (Supervisor Mode Execution Prevention)
2496 even if it is supported by processor.
2499 This affects only 32-bit executables.
2500 noexec32=on: enable non-executable mappings (default)
2501 read doesn't imply executable mappings
2502 noexec32=off: disable non-executable mappings
2503 read implies executable mappings
2505 nofpu [MIPS,SH] Disable hardware FPU at boot time.
2507 nofxsr [BUGS=X86-32] Disables x86 floating point extended
2508 register save and restore. The kernel will only save
2509 legacy floating-point registers on task switch.
2511 nohugeiomap [KNL,x86] Disable kernel huge I/O mappings.
2513 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
2514 (bounds check bypass). With this option data leaks are
2515 possible in the system.
2517 nospectre_v2 [X86,PPC_FSL_BOOK3E] Disable all mitigations for the Spectre variant 2
2518 (indirect branch prediction) vulnerability. System may
2519 allow data leaks with this option, which is equivalent
2522 nospec_store_bypass_disable
2523 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
2526 [PPC] Don't flush the L1-D cache after accessing user data.
2528 noxsave [BUGS=X86] Disables x86 extended register state save
2529 and restore using xsave. The kernel will fallback to
2530 enabling legacy floating-point and sse state.
2532 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
2533 register states. The kernel will fall back to use
2534 xsave to save the states. By using this parameter,
2535 performance of saving the states is degraded because
2536 xsave doesn't support modified optimization while
2537 xsaveopt supports it on xsaveopt enabled systems.
2539 noxsaves [X86] Disables xsaves and xrstors used in saving and
2540 restoring x86 extended register state in compacted
2541 form of xsave area. The kernel will fall back to use
2542 xsaveopt and xrstor to save and restore the states
2543 in standard form of xsave area. By using this
2544 parameter, xsave area per process might occupy more
2545 memory on xsaves enabled systems.
2547 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
2548 wfi(ARM) instruction doesn't work correctly and not to
2549 use it. This is also useful when using JTAG debugger.
2551 no_file_caps Tells the kernel not to honor file capabilities. The
2552 only way then for a file to be executed with privilege
2553 is to be setuid root or executed by root.
2555 nohalt [IA-64] Tells the kernel not to use the power saving
2556 function PAL_HALT_LIGHT when idle. This increases
2557 power-consumption. On the positive side, it reduces
2558 interrupt wake-up latency, which may improve performance
2559 in certain environments such as networked servers or
2562 nohibernate [HIBERNATION] Disable hibernation and resume.
2564 nohz= [KNL] Boottime enable/disable dynamic ticks
2565 Valid arguments: on, off
2568 nohz_full= [KNL,BOOT]
2569 In kernels built with CONFIG_NO_HZ_FULL=y, set
2570 the specified list of CPUs whose tick will be stopped
2571 whenever possible. The boot CPU will be forced outside
2572 the range to maintain the timekeeping.
2573 The CPUs in this range must also be included in the
2576 noiotrap [SH] Disables trapped I/O port accesses.
2578 noirqdebug [X86-32] Disables the code which attempts to detect and
2579 disable unhandled interrupt sources.
2581 no_timer_check [X86,APIC] Disables the code which tests for
2582 broken timer IRQ sources.
2584 noisapnp [ISAPNP] Disables ISA PnP code.
2586 noinitrd [RAM] Tells the kernel not to load any configured
2589 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
2591 [Deprecated - use intremap=off]
2595 noinvpcid [X86] Disable the INVPCID cpu feature.
2597 nojitter [IA-64] Disables jitter checking for ITC timers.
2599 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
2601 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
2604 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
2605 steal time is computed, but won't influence scheduler
2608 nolapic [X86-32,APIC] Do not enable or use the local APIC.
2610 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
2612 noltlbs [PPC] Do not use large page/tlb entries for kernel
2613 lowmem mapping on PPC40x.
2615 nomca [IA-64] Disable machine check abort handling
2617 nomce [X86-32] Disable Machine Check Exception
2619 nomfgpt [X86-32] Disable Multi-Function General Purpose
2620 Timer usage (for AMD Geode machines).
2622 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
2623 shutdown the other cpus. Instead use the REBOOT_VECTOR
2626 nomodule Disable module load
2628 nopat [X86] Disable PAT (page attribute table extension of
2629 pagetables) support.
2631 nopcid [X86-64] Disable the PCID cpu feature.
2633 norandmaps Don't use address space randomization. Equivalent to
2634 echo 0 > /proc/sys/kernel/randomize_va_space
2636 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
2637 with UP alternatives
2639 nordrand [X86] Disable kernel use of the RDRAND and
2640 RDSEED instructions even if they are supported
2641 by the processor. RDRAND and RDSEED are still
2642 available to user space applications.
2644 noresume [SWSUSP] Disables resume and restores original swap
2647 no-scroll [VGA] Disables scrollback.
2648 This is required for the Braillex ib80-piezo Braille
2649 reader made by F.H. Papenmeier (Germany).
2653 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
2655 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
2656 and disable the IO APIC. legacy for "maxcpus=0".
2658 nosoftlockup [KNL] Disable the soft-lockup detector.
2660 nosync [HW,M68K] Disables sync negotiation for all devices.
2662 notsc [BUGS=X86-32] Disable Time Stamp Counter
2664 nousb [USB] Disable the USB subsystem
2666 nowatchdog [KNL] Disable both lockup detectors, i.e.
2667 soft-lockup and NMI watchdog (hard-lockup).
2671 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
2673 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
2674 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
2675 Some features depend on CPU0. Known dependencies are:
2676 1. Resume from suspend/hibernate depends on CPU0.
2677 Suspend/hibernate will fail if CPU0 is offline and you
2678 need to online CPU0 before suspend/hibernate.
2679 2. PIC interrupts also depend on CPU0. CPU0 can't be
2680 removed if a PIC interrupt is detected.
2681 It's said poweroff/reboot may depend on CPU0 on some
2682 machines although I haven't seen such issues so far
2683 after CPU0 is offline on a few tested machines.
2684 If the dependencies are under your control, you can
2685 turn on cpu0_hotplug.
2687 nptcg= [IA-64] Override max number of concurrent global TLB
2688 purges which is reported from either PAL_VM_SUMMARY or
2691 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
2692 could support. nr_cpus=n : n >= 1 limits the kernel to
2693 supporting 'n' processors. Later in runtime you can not
2694 use hotplug cpu feature to put more cpu back to online.
2695 just like you compile the kernel NR_CPUS=n
2697 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
2699 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
2700 Allowed values are enable and disable
2702 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
2703 one of ['zone', 'node', 'default'] can be specified
2704 This can be set from sysctl after boot.
2705 See Documentation/sysctl/vm.txt for details.
2707 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
2708 See Documentation/debugging-via-ohci1394.txt for more
2711 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
2712 Rather than timing out after 20 ms if an EC
2713 command is not properly ACKed, override the length
2714 of the timeout. We have interrupts disabled while
2715 waiting for the ACK, so if this is set too high
2716 interrupts *may* be lost!
2718 omap_mux= [OMAP] Override bootloader pin multiplexing.
2719 Format: <mux_mode0.mode_name=value>...
2720 For example, to override I2C bus2:
2721 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
2723 oprofile.timer= [HW]
2724 Use timer interrupt instead of performance counters
2726 oprofile.cpu_type= Force an oprofile cpu type
2727 This might be useful if you have an older oprofile
2728 userland or if you want common events.
2729 Format: { arch_perfmon }
2730 arch_perfmon: [X86] Force use of architectural
2731 perfmon on Intel CPUs instead of the
2732 CPU specific event set.
2733 timer: [X86] Force use of architectural NMI
2734 timer mode (see also oprofile.timer
2735 for generic hr timer mode)
2736 [s390] Force legacy basic mode sampling
2737 (report cpu_type "timer")
2739 oops=panic Always panic on oopses. Default is to just kill the
2740 process, but there is a small probability of
2741 deadlocking the machine.
2742 This will also cause panics on machine check exceptions.
2743 Useful together with panic=30 to trigger a reboot.
2746 See Documentation/sound/oss/oss-parameters.txt
2748 page_owner= [KNL] Boot-time page_owner enabling option.
2749 Storage of the information about who allocated
2750 each page is disabled in default. With this switch,
2752 on: enable the feature
2754 panic= [KNL] Kernel behaviour on panic: delay <timeout>
2755 timeout > 0: seconds before rebooting
2756 timeout = 0: wait forever
2757 timeout < 0: reboot immediately
2760 panic_on_warn panic() instead of WARN(). Useful to cause kdump
2763 crash_kexec_post_notifiers
2764 Run kdump after running panic-notifiers and dumping
2765 kmsg. This only for the users who doubt kdump always
2766 succeeds in any situation.
2767 Note that this also increases risks of kdump failure,
2768 because some panic notifiers can make the crashed
2769 kernel more unstable.
2771 parkbd.port= [HW] Parallel port number the keyboard adapter is
2772 connected to, default is 0.
2774 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
2775 0 for XT, 1 for AT (default is AT).
2778 parport= [HW,PPT] Specify parallel ports. 0 disables.
2779 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
2780 Use 'auto' to force the driver to use any
2781 IRQ/DMA settings detected (the default is to
2782 ignore detected IRQ/DMA settings because of
2783 possible conflicts). You can specify the base
2784 address, IRQ, and DMA settings; IRQ and DMA
2785 should be numbers, or 'auto' (for using detected
2786 settings on that particular port), or 'nofifo'
2787 (to avoid using a FIFO even if it is detected).
2788 Parallel ports are assigned in the order they
2789 are specified on the command line, starting
2792 parport_init_mode= [HW,PPT]
2793 Configure VIA parallel port to operate in
2794 a specific mode. This is necessary on Pegasos
2795 computer where firmware has no options for setting
2796 up parallel port mode and sets it to spp.
2797 Currently this function knows 686a and 8231 chips.
2798 Format: [spp|ps2|epp|ecp|ecpepp]
2801 Halt all CPUs after the first oops has been printed for
2802 the specified number of seconds. This is to be used if
2803 your oopses keep scrolling off the screen.
2808 See header of drivers/block/paride/pcd.c.
2809 See also Documentation/blockdev/paride.txt.
2811 pci=option[,option...] [PCI] various PCI subsystem options:
2812 earlydump [X86] dump PCI config space before the kernel
2814 off [X86] don't probe for the PCI bus
2815 bios [X86-32] force use of PCI BIOS, don't access
2816 the hardware directly. Use this if your machine
2817 has a non-standard PCI host bridge.
2818 nobios [X86-32] disallow use of PCI BIOS, only direct
2819 hardware access methods are allowed. Use this
2820 if you experience crashes upon bootup and you
2821 suspect they are caused by the BIOS.
2822 conf1 [X86] Force use of PCI Configuration
2824 conf2 [X86] Force use of PCI Configuration
2826 noaer [PCIE] If the PCIEAER kernel config parameter is
2827 enabled, this kernel boot option can be used to
2828 disable the use of PCIE advanced error reporting.
2829 nodomains [PCI] Disable support for multiple PCI
2830 root domains (aka PCI segments, in ACPI-speak).
2831 nommconf [X86] Disable use of MMCONFIG for PCI
2833 check_enable_amd_mmconf [X86] check for and enable
2834 properly configured MMIO access to PCI
2835 config space on AMD family 10h CPU
2836 nomsi [MSI] If the PCI_MSI kernel config parameter is
2837 enabled, this kernel boot option can be used to
2838 disable the use of MSI interrupts system-wide.
2839 noioapicquirk [APIC] Disable all boot interrupt quirks.
2840 Safety option to keep boot IRQs enabled. This
2841 should never be necessary.
2842 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
2843 primary IO-APIC for bridges that cannot disable
2844 boot IRQs. This fixes a source of spurious IRQs
2845 when the system masks IRQs.
2846 noioapicreroute [APIC] Disable workaround that uses the
2847 boot IRQ equivalent of an IRQ that connects to
2848 a chipset where boot IRQs cannot be disabled.
2849 The opposite of ioapicreroute.
2850 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
2851 routing table. These calls are known to be buggy
2852 on several machines and they hang the machine
2853 when used, but on other computers it's the only
2854 way to get the interrupt routing table. Try
2855 this option if the kernel is unable to allocate
2856 IRQs or discover secondary PCI buses on your
2858 rom [X86] Assign address space to expansion ROMs.
2859 Use with caution as certain devices share
2860 address decoders between ROMs and other
2862 norom [X86] Do not assign address space to
2863 expansion ROMs that do not already have
2864 BIOS assigned address ranges.
2865 nobar [X86] Do not assign address space to the
2866 BARs that weren't assigned by the BIOS.
2867 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
2868 assigned automatically to PCI devices. You can
2869 make the kernel exclude IRQs of your ISA cards
2871 pirqaddr=0xAAAAA [X86] Specify the physical address
2872 of the PIRQ table (normally generated
2873 by the BIOS) if it is outside the
2874 F0000h-100000h range.
2875 lastbus=N [X86] Scan all buses thru bus #N. Can be
2876 useful if the kernel is unable to find your
2877 secondary buses and you want to tell it
2878 explicitly which ones they are.
2879 assign-busses [X86] Always assign all PCI bus
2880 numbers ourselves, overriding
2881 whatever the firmware may have done.
2882 usepirqmask [X86] Honor the possible IRQ mask stored
2883 in the BIOS $PIR table. This is needed on
2884 some systems with broken BIOSes, notably
2885 some HP Pavilion N5400 and Omnibook XE3
2886 notebooks. This will have no effect if ACPI
2887 IRQ routing is enabled.
2888 noacpi [X86] Do not use ACPI for IRQ routing
2889 or for PCI scanning.
2890 use_crs [X86] Use PCI host bridge window information
2891 from ACPI. On BIOSes from 2008 or later, this
2892 is enabled by default. If you need to use this,
2893 please report a bug.
2894 nocrs [X86] Ignore PCI host bridge windows from ACPI.
2895 If you need to use this, please report a bug.
2896 routeirq Do IRQ routing for all PCI devices.
2897 This is normally done in pci_enable_device(),
2898 so this option is a temporary workaround
2899 for broken drivers that don't call it.
2900 skip_isa_align [X86] do not align io start addr, so can
2901 handle more pci cards
2902 firmware [ARM] Do not re-enumerate the bus but instead
2903 just use the configuration from the
2904 bootloader. This is currently used on
2905 IXP2000 systems where the bus has to be
2906 configured a certain way for adjunct CPUs.
2907 noearly [X86] Don't do any early type 1 scanning.
2908 This might help on some broken boards which
2909 machine check when some devices' config space
2910 is read. But various workarounds are disabled
2911 and some IOMMU drivers will not work.
2912 bfsort Sort PCI devices into breadth-first order.
2913 This sorting is done to get a device
2914 order compatible with older (<= 2.4) kernels.
2915 nobfsort Don't sort PCI devices into breadth-first order.
2916 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
2917 tuning and use the BIOS-configured MPS defaults.
2918 pcie_bus_safe Set every device's MPS to the largest value
2919 supported by all devices below the root complex.
2920 pcie_bus_perf Set device MPS to the largest allowable MPS
2921 based on its parent bus. Also set MRRS (Max
2922 Read Request Size) to the largest supported
2923 value (no larger than the MPS that the device
2924 or bus can support) for best performance.
2925 pcie_bus_peer2peer Set every device's MPS to 128B, which
2926 every device is guaranteed to support. This
2927 configuration allows peer-to-peer DMA between
2928 any pair of devices, possibly at the cost of
2929 reduced performance. This also guarantees
2930 that hot-added devices will work.
2931 cbiosize=nn[KMG] The fixed amount of bus space which is
2932 reserved for the CardBus bridge's IO window.
2933 The default value is 256 bytes.
2934 cbmemsize=nn[KMG] The fixed amount of bus space which is
2935 reserved for the CardBus bridge's memory
2936 window. The default value is 64 megabytes.
2939 [<order of align>@][<domain>:]<bus>:<slot>.<func>[; ...]
2940 Specifies alignment and device to reassign
2941 aligned memory resources.
2942 If <order of align> is not specified,
2943 PAGE_SIZE is used as alignment.
2944 PCI-PCI bridge can be specified, if resource
2945 windows need to be expanded.
2946 ecrc= Enable/disable PCIe ECRC (transaction layer
2947 end-to-end CRC checking).
2948 bios: Use BIOS/firmware settings. This is the
2952 hpiosize=nn[KMG] The fixed amount of bus space which is
2953 reserved for hotplug bridge's IO window.
2954 Default size is 256 bytes.
2955 hpmemsize=nn[KMG] The fixed amount of bus space which is
2956 reserved for hotplug bridge's memory window.
2957 Default size is 2 megabytes.
2958 realloc= Enable/disable reallocating PCI bridge resources
2959 if allocations done by BIOS are too small to
2960 accommodate resources required by all child
2962 off: Turn realloc off
2964 realloc same as realloc=on
2965 noari do not use PCIe ARI.
2966 pcie_scan_all Scan all possible PCIe devices. Otherwise we
2967 only look for one device below a PCIe downstream
2970 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
2973 force Enable ASPM even on devices that claim not to support it.
2974 WARNING: Forcing ASPM on may cause system lockups.
2976 pcie_hp= [PCIE] PCI Express Hotplug driver options:
2977 nomsi Do not use MSI for PCI Express Native Hotplug (this
2978 makes all PCIe ports use INTx for hotplug services).
2980 pcie_ports= [PCIE] PCIe ports handling:
2981 auto Ask the BIOS whether or not to use native PCIe services
2982 associated with PCIe ports (PME, hot-plug, AER). Use
2983 them only if that is allowed by the BIOS.
2984 native Use native PCIe services associated with PCIe ports
2986 compat Treat PCIe ports as PCI-to-PCI bridges, disable the PCIe
2989 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
2990 nomsi Do not use MSI for native PCIe PME signaling (this makes
2991 all PCIe root ports use INTx for all services).
2993 pcmv= [HW,PCMCIA] BadgePAD 4
2997 Keep all power-domains already enabled by bootloader on,
2998 even if no driver has claimed them. This is useful
2999 for debug and development, but should not be
3000 needed on a platform with proper driver support.
3003 See Documentation/blockdev/paride.txt.
3005 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3008 See arch/parisc/kernel/pdc_chassis.c
3010 percpu_alloc= Select which percpu first chunk allocator to use.
3011 Currently supported values are "embed" and "page".
3012 Archs may support subset or none of the selections.
3013 See comments in mm/percpu.c for details on each
3014 allocator. This parameter is primarily for debugging
3015 and performance comparison.
3018 See Documentation/blockdev/paride.txt.
3021 See Documentation/blockdev/paride.txt.
3023 pirq= [SMP,APIC] Manual mp-table setup
3024 See Documentation/x86/i386/IO-APIC.txt.
3026 plip= [PPT,NET] Parallel port network link
3027 Format: { parport<nr> | timid | 0 }
3028 See also Documentation/parport.txt.
3030 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3031 Override pmtimer IOPort with a hex value.
3035 Enable PNP debug messages (depends on the
3036 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3037 via /sys/module/pnp/parameters/debug. We always show
3038 current resource usage; turning this on also shows
3039 possible settings and some assignment information.
3045 { on | off | curr | res | no-curr | no-res }
3048 [ISAPNP] Exclude IRQs for the autoconfiguration
3051 [ISAPNP] Exclude DMAs for the autoconfiguration
3053 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3054 Ranges are in pairs (I/O port base and size).
3057 [ISAPNP] Exclude memory regions for the
3059 Ranges are in pairs (memory base and size).
3061 ports= [IP_VS_FTP] IPVS ftp helper module
3063 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3065 Format: <port>,<port>....
3067 print-fatal-signals=
3068 [KNL] debug: print fatal signals
3070 If enabled, warn about various signal handling
3071 related application anomalies: too many signals,
3072 too many POSIX.1 timers, fatal signals causing a
3075 If you hit the warning due to signal overflow,
3076 you might want to try "ulimit -i unlimited".
3080 printk.always_kmsg_dump=
3081 Trigger kmsg_dump for cases other than kernel oops or
3083 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3086 printk.time= Show timing data prefixed to each printk message line
3087 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3089 processor.max_cstate= [HW,ACPI]
3090 Limit processor to maximum C-state
3091 max_cstate=9 overrides any DMI blacklist limit.
3093 processor.nocst [HW,ACPI]
3094 Ignore the _CST method to determine C-states,
3095 instead using the legacy FADT method
3097 profile= [KNL] Enable kernel profiling via /proc/profile
3098 Format: [schedule,]<number>
3099 Param: "schedule" - profile schedule points.
3100 Param: <number> - step/bucket size as a power of 2 for
3101 statistical time based profiling.
3102 Param: "sleep" - profile D-state sleeping (millisecs).
3103 Requires CONFIG_SCHEDSTATS
3104 Param: "kvm" - profile VM exits.
3106 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3108 See Documentation/blockdev/ramdisk.txt.
3110 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3111 probe for; one of (bare|imps|exps|lifebook|any).
3112 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3114 psmouse.resetafter= [HW,MOUSE]
3115 Try to reset the device after so many bad packets
3118 [HW,MOUSE] Set desired mouse resolution, in dpi.
3119 psmouse.smartscroll=
3120 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3121 0 = disabled, 1 = enabled (default).
3123 pstore.backend= Specify the name of the pstore backend to use
3126 See Documentation/blockdev/paride.txt.
3128 pti= [X86_64] Control Page Table Isolation of user and
3129 kernel address spaces. Disabling this feature
3130 removes hardening, but improves performance of
3131 system calls and interrupts.
3133 on - unconditionally enable
3134 off - unconditionally disable
3135 auto - kernel detects whether your CPU model is
3136 vulnerable to issues that PTI mitigates
3138 Not specifying this option is equivalent to pti=auto.
3141 Equivalent to pti=off
3144 [KNL] Number of legacy pty's. Overwrites compiled-in
3147 quiet [KNL] Disable most log messages
3152 See Documentation/md.txt.
3154 ramdisk_blocksize= [RAM]
3155 See Documentation/blockdev/ramdisk.txt.
3157 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3158 See Documentation/blockdev/ramdisk.txt.
3161 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3162 the specified list of CPUs to be no-callback CPUs.
3163 Invocation of these CPUs' RCU callbacks will
3164 be offloaded to "rcuox/N" kthreads created for
3165 that purpose, where "x" is "b" for RCU-bh, "p"
3166 for RCU-preempt, and "s" for RCU-sched, and "N"
3167 is the CPU number. This reduces OS jitter on the
3168 offloaded CPUs, which can be useful for HPC and
3169 real-time workloads. It can also improve energy
3170 efficiency for asymmetric multiprocessors.
3173 Rather than requiring that offloaded CPUs
3174 (specified by rcu_nocbs= above) explicitly
3175 awaken the corresponding "rcuoN" kthreads,
3176 make these kthreads poll for callbacks.
3177 This improves the real-time response for the
3178 offloaded CPUs by relieving them of the need to
3179 wake up the corresponding kthread, but degrades
3180 energy efficiency by requiring that the kthreads
3181 periodically wake up to do the polling.
3183 rcutree.blimit= [KNL]
3184 Set maximum number of finished RCU callbacks to
3185 process in one batch.
3187 rcutree.dump_tree= [KNL]
3188 Dump the structure of the rcu_node combining tree
3189 out at early boot. This is used for diagnostic
3190 purposes, to verify correct tree setup.
3192 rcutree.gp_cleanup_delay= [KNL]
3193 Set the number of jiffies to delay each step of
3194 RCU grace-period cleanup. This only has effect
3195 when CONFIG_RCU_TORTURE_TEST_SLOW_CLEANUP is set.
3197 rcutree.gp_init_delay= [KNL]
3198 Set the number of jiffies to delay each step of
3199 RCU grace-period initialization. This only has
3200 effect when CONFIG_RCU_TORTURE_TEST_SLOW_INIT
3203 rcutree.gp_preinit_delay= [KNL]
3204 Set the number of jiffies to delay each step of
3205 RCU grace-period pre-initialization, that is,
3206 the propagation of recent CPU-hotplug changes up
3207 the rcu_node combining tree. This only has effect
3208 when CONFIG_RCU_TORTURE_TEST_SLOW_PREINIT is set.
3210 rcutree.rcu_fanout_exact= [KNL]
3211 Disable autobalancing of the rcu_node combining
3212 tree. This is used by rcutorture, and might
3213 possibly be useful for architectures having high
3214 cache-to-cache transfer latencies.
3216 rcutree.rcu_fanout_leaf= [KNL]
3217 Change the number of CPUs assigned to each
3218 leaf rcu_node structure. Useful for very
3219 large systems, which will choose the value 64,
3220 and for NUMA systems with large remote-access
3221 latencies, which will choose a value aligned
3222 with the appropriate hardware boundaries.
3224 rcutree.jiffies_till_sched_qs= [KNL]
3225 Set required age in jiffies for a
3226 given grace period before RCU starts
3227 soliciting quiescent-state help from
3228 rcu_note_context_switch().
3230 rcutree.jiffies_till_first_fqs= [KNL]
3231 Set delay from grace-period initialization to
3232 first attempt to force quiescent states.
3233 Units are jiffies, minimum value is zero,
3234 and maximum value is HZ.
3236 rcutree.jiffies_till_next_fqs= [KNL]
3237 Set delay between subsequent attempts to force
3238 quiescent states. Units are jiffies, minimum
3239 value is one, and maximum value is HZ.
3241 rcutree.kthread_prio= [KNL,BOOT]
3242 Set the SCHED_FIFO priority of the RCU per-CPU
3243 kthreads (rcuc/N). This value is also used for
3244 the priority of the RCU boost threads (rcub/N)
3245 and for the RCU grace-period kthreads (rcu_bh,
3246 rcu_preempt, and rcu_sched). If RCU_BOOST is
3247 set, valid values are 1-99 and the default is 1
3248 (the least-favored priority). Otherwise, when
3249 RCU_BOOST is not set, valid values are 0-99 and
3250 the default is zero (non-realtime operation).
3252 rcutree.rcu_nocb_leader_stride= [KNL]
3253 Set the number of NOCB kthread groups, which
3254 defaults to the square root of the number of
3255 CPUs. Larger numbers reduces the wakeup overhead
3256 on the per-CPU grace-period kthreads, but increases
3257 that same overhead on each group's leader.
3259 rcutree.qhimark= [KNL]
3260 Set threshold of queued RCU callbacks beyond which
3261 batch limiting is disabled.
3263 rcutree.qlowmark= [KNL]
3264 Set threshold of queued RCU callbacks below which
3265 batch limiting is re-enabled.
3267 rcutree.rcu_idle_gp_delay= [KNL]
3268 Set wakeup interval for idle CPUs that have
3269 RCU callbacks (RCU_FAST_NO_HZ=y).
3271 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3272 Set wakeup interval for idle CPUs that have
3273 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3274 Lazy RCU callbacks are those which RCU can
3275 prove do nothing more than free memory.
3277 rcutorture.cbflood_inter_holdoff= [KNL]
3278 Set holdoff time (jiffies) between successive
3279 callback-flood tests.
3281 rcutorture.cbflood_intra_holdoff= [KNL]
3282 Set holdoff time (jiffies) between successive
3283 bursts of callbacks within a given callback-flood
3286 rcutorture.cbflood_n_burst= [KNL]
3287 Set the number of bursts making up a given
3288 callback-flood test. Set this to zero to
3289 disable callback-flood testing.
3291 rcutorture.cbflood_n_per_burst= [KNL]
3292 Set the number of callbacks to be registered
3293 in a given burst of a callback-flood test.
3295 rcutorture.fqs_duration= [KNL]
3296 Set duration of force_quiescent_state bursts
3299 rcutorture.fqs_holdoff= [KNL]
3300 Set holdoff time within force_quiescent_state bursts
3303 rcutorture.fqs_stutter= [KNL]
3304 Set wait time between force_quiescent_state bursts
3307 rcutorture.gp_cond= [KNL]
3308 Use conditional/asynchronous update-side
3309 primitives, if available.
3311 rcutorture.gp_exp= [KNL]
3312 Use expedited update-side primitives, if available.
3314 rcutorture.gp_normal= [KNL]
3315 Use normal (non-expedited) asynchronous
3316 update-side primitives, if available.
3318 rcutorture.gp_sync= [KNL]
3319 Use normal (non-expedited) synchronous
3320 update-side primitives, if available. If all
3321 of rcutorture.gp_cond=, rcutorture.gp_exp=,
3322 rcutorture.gp_normal=, and rcutorture.gp_sync=
3323 are zero, rcutorture acts as if is interpreted
3324 they are all non-zero.
3326 rcutorture.n_barrier_cbs= [KNL]
3327 Set callbacks/threads for rcu_barrier() testing.
3329 rcutorture.nfakewriters= [KNL]
3330 Set number of concurrent RCU writers. These just
3331 stress RCU, they don't participate in the actual
3332 test, hence the "fake".
3334 rcutorture.nreaders= [KNL]
3335 Set number of RCU readers. The value -1 selects
3336 N-1, where N is the number of CPUs. A value
3337 "n" less than -1 selects N-n-2, where N is again
3338 the number of CPUs. For example, -2 selects N
3339 (the number of CPUs), -3 selects N+1, and so on.
3341 rcutorture.object_debug= [KNL]
3342 Enable debug-object double-call_rcu() testing.
3344 rcutorture.onoff_holdoff= [KNL]
3345 Set time (s) after boot for CPU-hotplug testing.
3347 rcutorture.onoff_interval= [KNL]
3348 Set time (s) between CPU-hotplug operations, or
3349 zero to disable CPU-hotplug testing.
3351 rcutorture.shuffle_interval= [KNL]
3352 Set task-shuffle interval (s). Shuffling tasks
3353 allows some CPUs to go into dyntick-idle mode
3354 during the rcutorture test.
3356 rcutorture.shutdown_secs= [KNL]
3357 Set time (s) after boot system shutdown. This
3358 is useful for hands-off automated testing.
3360 rcutorture.stall_cpu= [KNL]
3361 Duration of CPU stall (s) to test RCU CPU stall
3362 warnings, zero to disable.
3364 rcutorture.stall_cpu_holdoff= [KNL]
3365 Time to wait (s) after boot before inducing stall.
3367 rcutorture.stat_interval= [KNL]
3368 Time (s) between statistics printk()s.
3370 rcutorture.stutter= [KNL]
3371 Time (s) to stutter testing, for example, specifying
3372 five seconds causes the test to run for five seconds,
3373 wait for five seconds, and so on. This tests RCU's
3374 ability to transition abruptly to and from idle.
3376 rcutorture.test_boost= [KNL]
3377 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
3378 "Maybe" means test if the RCU implementation
3379 under test support RCU priority boosting.
3381 rcutorture.test_boost_duration= [KNL]
3382 Duration (s) of each individual boost test.
3384 rcutorture.test_boost_interval= [KNL]
3385 Interval (s) between each boost test.
3387 rcutorture.test_no_idle_hz= [KNL]
3388 Test RCU's dyntick-idle handling. See also the
3389 rcutorture.shuffle_interval parameter.
3391 rcutorture.torture_runnable= [BOOT]
3392 Start rcutorture running at boot time.
3394 rcutorture.torture_type= [KNL]
3395 Specify the RCU implementation to test.
3397 rcutorture.verbose= [KNL]
3398 Enable additional printk() statements.
3400 rcupdate.rcu_expedited= [KNL]
3401 Use expedited grace-period primitives, for
3402 example, synchronize_rcu_expedited() instead
3403 of synchronize_rcu(). This reduces latency,
3404 but can increase CPU utilization, degrade
3405 real-time latency, and degrade energy efficiency.
3407 rcupdate.rcu_cpu_stall_suppress= [KNL]
3408 Suppress RCU CPU stall warning messages.
3410 rcupdate.rcu_cpu_stall_timeout= [KNL]
3411 Set timeout for RCU CPU stall warning messages.
3413 rcupdate.rcu_task_stall_timeout= [KNL]
3414 Set timeout in jiffies for RCU task stall warning
3415 messages. Disable with a value less than or equal
3418 rcupdate.rcu_self_test= [KNL]
3419 Run the RCU early boot self tests
3421 rcupdate.rcu_self_test_bh= [KNL]
3422 Run the RCU bh early boot self tests
3424 rcupdate.rcu_self_test_sched= [KNL]
3425 Run the RCU sched early boot self tests
3429 Run specified binary instead of /init from the ramdisk,
3430 used for early userspace startup. See initrd.
3433 force - Override the decision by the kernel to hide the
3434 advertisement of RDRAND support (this affects
3435 certain AMD processors because of buggy BIOS
3436 support, specifically around the suspend/resume
3440 Format (x86 or x86_64):
3441 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
3443 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
3445 Where reboot_mode is one of warm (soft) or cold (hard) or gpio,
3446 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
3447 reboot_force is either force or not specified,
3448 reboot_cpu is s[mp]#### with #### being the processor
3449 to be used for rebooting.
3452 [KNL, SMP] Set scheduler's default relax_domain_level.
3453 See Documentation/cgroups/cpusets.txt.
3455 relative_sleep_states=
3456 [SUSPEND] Use sleep state labeling where the deepest
3457 state available other than hibernation is always "mem".
3458 Format: { "0" | "1" }
3459 0 -- Traditional sleep state labels.
3460 1 -- Relative sleep state labels.
3462 reserve= [KNL,BUGS] Force the kernel to ignore some iomem area
3464 reservetop= [X86-32]
3466 Reserves a hole at the top of the kernel virtual
3471 Set the amount of memory to reserve for BIOS at
3472 the bottom of the address space.
3474 reset_devices [KNL] Force drivers to reset the underlying device
3475 during initialization.
3478 Specify the partition device for software suspend
3480 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
3482 resume_offset= [SWSUSP]
3483 Specify the offset from the beginning of the partition
3484 given by "resume=" at which the swap header is located,
3485 in <PAGE_SIZE> units (needed only for swap files).
3486 See Documentation/power/swsusp-and-swap-files.txt
3488 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
3489 read the resume files
3491 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
3492 Useful for devices that are detected asynchronously
3493 (e.g. USB and MMC devices).
3495 hibernate= [HIBERNATION]
3496 noresume Don't check if there's a hibernation image
3497 present during boot.
3498 nocompress Don't compress/decompress hibernation images.
3499 no Disable hibernation and resume.
3501 retain_initrd [RAM] Keep initrd memory after extraction
3503 rfkill.default_state=
3504 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
3505 etc. communication is blocked by default.
3508 rfkill.master_switch_mode=
3509 0 The "airplane mode" button does nothing.
3510 1 The "airplane mode" button toggles between everything
3511 blocked and the previous configuration.
3512 2 The "airplane mode" button toggles between everything
3513 blocked and everything unblocked.
3515 rhash_entries= [KNL,NET]
3516 Set number of hash buckets for route cache
3518 ro [KNL] Mount root device read-only on boot
3520 root= [KNL] Root filesystem
3521 See name_to_dev_t comment in init/do_mounts.c.
3523 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
3524 mount the root filesystem
3526 rootflags= [KNL] Set root filesystem mount option string
3528 rootfstype= [KNL] Set root filesystem type
3530 rootwait [KNL] Wait (indefinitely) for root device to show up.
3531 Useful for devices that are detected asynchronously
3532 (e.g. USB and MMC devices).
3534 rproc_mem=nn[KMG][@address]
3535 [KNL,ARM,CMA] Remoteproc physical memory block.
3536 Memory area to be used by remote processor image,
3539 rw [KNL] Mount root device read-write on boot
3541 S [KNL] Run init in single mode
3543 s390_iommu= [HW,S390]
3544 Set s390 IOTLB flushing mode
3546 With strict flushing every unmap operation will result in
3547 an IOTLB flush. Default is lazy flushing before reuse,
3551 See drivers/net/irda/sa1100_ir.c.
3553 sbni= [NET] Granch SBNI12 leased line adapter
3555 sched_debug [KNL] Enables verbose scheduler debug messages.
3557 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
3558 xtime_lock contention on larger systems, and/or RCU lock
3559 contention on all systems with CONFIG_MAXSMP set.
3560 Format: { "0" | "1" }
3561 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
3563 Note: increases power consumption, thus should only be
3564 enabled if running jitter sensitive (HPC/RT) workloads.
3566 security= [SECURITY] Choose a security module to enable at boot.
3567 If this boot parameter is not specified, only the first
3568 security module asking for security registration will be
3569 loaded. An invalid security module name will be treated
3570 as if no module has been chosen.
3572 selinux= [SELINUX] Disable or enable SELinux at boot time.
3573 Format: { "0" | "1" }
3574 See security/selinux/Kconfig help text.
3577 Default value is set via kernel config option.
3578 If enabled at boot time, /selinux/disable can be used
3579 later to disable prior to initial policy load.
3581 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
3582 Format: { "0" | "1" }
3583 See security/apparmor/Kconfig help text
3586 Default value is set via kernel config option.
3588 serialnumber [BUGS=X86-32]
3591 Maximal number of shapers.
3593 show_msr= [x86] show boot-time MSR settings
3594 Format: { <integer> }
3595 Show boot-time (BIOS-initialized) MSR settings.
3596 The parameter means the number of CPUs to show,
3597 for example 1 means boot CPU only.
3605 Disable merging of slabs with similar size. May be
3606 necessary if there is some reason to distinguish
3607 allocs to different slabs. Debug options disable
3608 merging on their own.
3609 For more information see Documentation/vm/slub.txt.
3611 slab_max_order= [MM, SLAB]
3612 Determines the maximum allowed order for slabs.
3613 A high setting may cause OOMs due to memory
3614 fragmentation. Defaults to 1 for systems with
3615 more than 32MB of RAM, 0 otherwise.
3617 slub_debug[=options[,slabs]] [MM, SLUB]
3618 Enabling slub_debug allows one to determine the
3619 culprit if slab objects become corrupted. Enabling
3620 slub_debug can create guard zones around objects and
3621 may poison objects when not in use. Also tracks the
3622 last alloc / free. For more information see
3623 Documentation/vm/slub.txt.
3625 slub_max_order= [MM, SLUB]
3626 Determines the maximum allowed order for slabs.
3627 A high setting may cause OOMs due to memory
3628 fragmentation. For more information see
3629 Documentation/vm/slub.txt.
3631 slub_min_objects= [MM, SLUB]
3632 The minimum number of objects per slab. SLUB will
3633 increase the slab order up to slub_max_order to
3634 generate a sufficiently large slab able to contain
3635 the number of objects indicated. The higher the number
3636 of objects the smaller the overhead of tracking slabs
3637 and the less frequently locks need to be acquired.
3638 For more information see Documentation/vm/slub.txt.
3640 slub_min_order= [MM, SLUB]
3641 Determines the minimum page order for slabs. Must be
3642 lower than slub_max_order.
3643 For more information see Documentation/vm/slub.txt.
3645 slub_nomerge [MM, SLUB]
3646 Same with slab_nomerge. This is supported for legacy.
3647 See slab_nomerge for more information.
3650 Format: <io1>[,<io2>[,...,<io8>]]
3652 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
3653 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
3654 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
3655 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
3656 smsc-ircc2.ircc_irq= [HW] IRQ line
3657 smsc-ircc2.ircc_dma= [HW] DMA channel
3658 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
3659 0: Toshiba Satellite 1800 (GP data pin select)
3660 1: Fast pin select (default)
3664 [KNL] Should the soft-lockup detector generate panics.
3667 softlockup_all_cpu_backtrace=
3668 [KNL] Should the soft-lockup detector generate
3669 backtraces on all cpus.
3672 sonypi.*= [HW] Sony Programmable I/O Control Device driver
3673 See Documentation/laptops/sonypi.txt
3675 spectre_v2= [X86] Control mitigation of Spectre variant 2
3676 (indirect branch speculation) vulnerability.
3677 The default operation protects the kernel from
3680 on - unconditionally enable, implies
3682 off - unconditionally disable, implies
3684 auto - kernel detects whether your CPU model is
3687 Selecting 'on' will, and 'auto' may, choose a
3688 mitigation method at run time according to the
3689 CPU, the available microcode, the setting of the
3690 CONFIG_RETPOLINE configuration option, and the
3691 compiler with which the kernel was built.
3693 Selecting 'on' will also enable the mitigation
3694 against user space to user space task attacks.
3696 Selecting 'off' will disable both the kernel and
3697 the user space protections.
3699 Specific mitigations can also be selected manually:
3701 retpoline - replace indirect branches
3702 retpoline,generic - google's original retpoline
3703 retpoline,amd - AMD-specific minimal thunk
3705 Not specifying this option is equivalent to
3709 [X86] Control mitigation of Spectre variant 2
3710 (indirect branch speculation) vulnerability between
3713 on - Unconditionally enable mitigations. Is
3714 enforced by spectre_v2=on
3716 off - Unconditionally disable mitigations. Is
3717 enforced by spectre_v2=off
3719 prctl - Indirect branch speculation is enabled,
3720 but mitigation can be enabled via prctl
3721 per thread. The mitigation control state
3722 is inherited on fork.
3725 - Like "prctl" above, but only STIBP is
3726 controlled per thread. IBPB is issued
3727 always when switching between different user
3731 - Same as "prctl" above, but all seccomp
3732 threads will enable the mitigation unless
3733 they explicitly opt out.
3736 - Like "seccomp" above, but only STIBP is
3737 controlled per thread. IBPB is issued
3738 always when switching between different
3739 user space processes.
3741 auto - Kernel selects the mitigation depending on
3742 the available CPU features and vulnerability.
3745 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
3747 Not specifying this option is equivalent to
3748 spectre_v2_user=auto.
3750 spec_store_bypass_disable=
3751 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
3752 (Speculative Store Bypass vulnerability)
3754 Certain CPUs are vulnerable to an exploit against a
3755 a common industry wide performance optimization known
3756 as "Speculative Store Bypass" in which recent stores
3757 to the same memory location may not be observed by
3758 later loads during speculative execution. The idea
3759 is that such stores are unlikely and that they can
3760 be detected prior to instruction retirement at the
3761 end of a particular speculation execution window.
3763 In vulnerable processors, the speculatively forwarded
3764 store can be used in a cache side channel attack, for
3765 example to read memory to which the attacker does not
3766 directly have access (e.g. inside sandboxed code).
3768 This parameter controls whether the Speculative Store
3769 Bypass optimization is used.
3771 on - Unconditionally disable Speculative Store Bypass
3772 off - Unconditionally enable Speculative Store Bypass
3773 auto - Kernel detects whether the CPU model contains an
3774 implementation of Speculative Store Bypass and
3775 picks the most appropriate mitigation. If the
3776 CPU is not vulnerable, "off" is selected. If the
3777 CPU is vulnerable the default mitigation is
3778 architecture and Kconfig dependent. See below.
3779 prctl - Control Speculative Store Bypass per thread
3780 via prctl. Speculative Store Bypass is enabled
3781 for a process by default. The state of the control
3782 is inherited on fork.
3783 seccomp - Same as "prctl" above, but all seccomp threads
3784 will disable SSB unless they explicitly opt out.
3786 Not specifying this option is equivalent to
3787 spec_store_bypass_disable=auto.
3789 Default mitigations:
3790 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
3792 spia_io_base= [HW,MTD]
3798 Control the Special Register Buffer Data Sampling
3801 Certain CPUs are vulnerable to an MDS-like
3802 exploit which can leak bits from the random
3805 By default, this issue is mitigated by
3806 microcode. However, the microcode fix can cause
3807 the RDRAND and RDSEED instructions to become
3808 much slower. Among other effects, this will
3809 result in reduced throughput from /dev/urandom.
3811 The microcode mitigation can be disabled with
3812 the following option:
3814 off: Disable mitigation and remove
3815 performance impact to RDRAND and RDSEED
3817 stack_guard_gap= [MM]
3818 override the default stack gap protection. The value
3819 is in page units and it defines how many pages prior
3820 to (for stacks growing down) resp. after (for stacks
3821 growing up) the main stack are reserved for no other
3822 mapping. Default value is 256 pages.
3825 Enabled the stack tracer on boot up.
3827 stacktrace_filter=[function-list]
3828 [FTRACE] Limit the functions that the stack tracer
3829 will trace at boot up. function-list is a comma separated
3830 list of functions. This list can be changed at run
3831 time by the stack_trace_filter file in the debugfs
3832 tracing directory. Note, this enables stack tracing
3833 and the stacktrace above is not needed.
3837 Set the STI (builtin display/keyboard on the HP-PARISC
3838 machines) console (graphic card) which should be used
3839 as the initial boot-console.
3840 See also comment in drivers/video/console/sticore.c.
3843 See comment in drivers/video/console/sticore.c.
3846 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
3848 sunrpc.min_resvport=
3849 sunrpc.max_resvport=
3851 SunRPC servers often require that client requests
3852 originate from a privileged port (i.e. a port in the
3853 range 0 < portnr < 1024).
3854 An administrator who wishes to reserve some of these
3855 ports for other uses may adjust the range that the
3856 kernel's sunrpc client considers to be privileged
3857 using these two parameters to set the minimum and
3858 maximum port values.
3862 Control how the NFS server code allocates CPUs to
3863 service thread pools. Depending on how many NICs
3864 you have and where their interrupts are bound, this
3865 option will affect which CPUs will do NFS serving.
3866 Note: this parameter cannot be changed while the
3867 NFS server is running.
3869 auto the server chooses an appropriate mode
3870 automatically using heuristics
3871 global a single global pool contains all CPUs
3872 percpu one pool for each CPU
3873 pernode one pool for each NUMA node (equivalent
3874 to global on non-NUMA machines)
3876 sunrpc.tcp_slot_table_entries=
3877 sunrpc.udp_slot_table_entries=
3879 Sets the upper limit on the number of simultaneous
3880 RPC calls that can be sent from the client to a
3881 server. Increasing these values may allow you to
3882 improve throughput, but will also increase the
3883 amount of memory reserved for use by the client.
3885 suspend.pm_test_delay=
3887 Sets the number of seconds to remain in a suspend test
3888 mode before resuming the system (see
3889 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
3890 is set. Default value is 5.
3893 [KNL] Enable accounting of swap in memory resource
3894 controller if no parameter or 1 is given or disable
3895 it if 0 is given (See Documentation/cgroups/memory.txt)
3897 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
3898 Format: { <int> | force }
3899 <int> -- Number of I/O TLB slabs
3900 force -- force using of bounce buffers even if they
3901 wouldn't be automatically used by the kernel
3905 sysfs.deprecated=0|1 [KNL]
3906 Enable/disable old style sysfs layout for old udev
3907 on older distributions. When this option is enabled
3908 very new udev will not work anymore. When this option
3909 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
3910 in older udev will not work anymore.
3911 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
3912 the kernel configuration.
3914 sysrq_always_enabled
3916 Ignore sysrq setting - this boot parameter will
3917 neutralize any effect of /proc/sys/kernel/sysrq.
3918 Useful for debugging.
3920 tcpmhash_entries= [KNL,NET]
3921 Set the number of tcp_metrics_hash slots.
3922 Default value is 8192 or 16384 depending on total
3923 ram pages. This is used to specify the TCP metrics
3924 cache size. See Documentation/networking/ip-sysctl.txt
3925 "tcp_no_metrics_save" section for more details.
3929 test_suspend= [SUSPEND][,N]
3930 Specify "mem" (for Suspend-to-RAM) or "standby" (for
3931 standby suspend) or "freeze" (for suspend type freeze)
3932 as the system sleep state during system startup with
3933 the optional capability to repeat N number of times.
3934 The system is woken from this state using a
3935 wakeup-capable RTC alarm.
3937 thash_entries= [KNL,NET]
3938 Set number of hash buckets for TCP connection
3940 thermal.act= [HW,ACPI]
3941 -1: disable all active trip points in all thermal zones
3942 <degrees C>: override all lowest active trip points
3944 thermal.crt= [HW,ACPI]
3945 -1: disable all critical trip points in all thermal zones
3946 <degrees C>: override all critical trip points
3948 thermal.nocrt= [HW,ACPI]
3949 Set to disable actions on ACPI thermal zone
3950 critical and hot trip points.
3952 thermal.off= [HW,ACPI]
3953 1: disable ACPI thermal control
3955 thermal.psv= [HW,ACPI]
3956 -1: disable all passive trip points
3957 <degrees C>: override all passive trip points to this
3960 thermal.tzp= [HW,ACPI]
3961 Specify global default ACPI thermal zone polling rate
3962 <deci-seconds>: poll all this frequency
3963 0: no polling (default)
3966 Force threading of all interrupt handlers except those
3967 marked explicitly IRQF_NO_THREAD.
3970 Enable the Transcendent memory driver if built-in.
3972 tmem.cleancache=0|1 [KNL, XEN]
3973 Default is on (1). Disable the usage of the cleancache
3974 API to send anonymous pages to the hypervisor.
3976 tmem.frontswap=0|1 [KNL, XEN]
3977 Default is on (1). Disable the usage of the frontswap
3978 API to send swap pages to the hypervisor. If disabled
3979 the selfballooning and selfshrinking are force disabled.
3981 tmem.selfballooning=0|1 [KNL, XEN]
3982 Default is on (1). Disable the driving of swap pages
3985 tmem.selfshrinking=0|1 [KNL, XEN]
3986 Default is on (1). Partial swapoff that immediately
3987 transfers pages from Xen hypervisor back to the
3988 kernel based on different criteria.
3992 Specify if the kernel should make use of the cpu
3993 topology information if the hardware supports this.
3994 The scheduler will make use of this information and
3995 e.g. base its process migration decisions on it.
3998 topology_updates= [KNL, PPC, NUMA]
4000 Specify if the kernel should ignore (off)
4001 topology updates sent by the hypervisor to this
4006 tpm_suspend_pcr=[HW,TPM]
4007 Format: integer pcr id
4008 Specify that at suspend time, the tpm driver
4009 should extend the specified pcr with zeros,
4010 as a workaround for some chips which fail to
4011 flush the last written pcr on TPM_SaveState.
4012 This will guarantee that all the other pcrs
4015 trace_buf_size=nn[KMG]
4016 [FTRACE] will set tracing buffer size on each cpu.
4018 trace_event=[event-list]
4019 [FTRACE] Set and start specified trace events in order
4020 to facilitate early boot debugging.
4021 See also Documentation/trace/events.txt
4023 trace_options=[option-list]
4024 [FTRACE] Enable or disable tracer options at boot.
4025 The option-list is a comma delimited list of options
4026 that can be enabled or disabled just as if you were
4027 to echo the option name into
4029 /sys/kernel/debug/tracing/trace_options
4031 For example, to enable stacktrace option (to dump the
4032 stack trace of each event), add to the command line:
4034 trace_options=stacktrace
4036 See also Documentation/trace/ftrace.txt "trace options"
4040 Have the tracepoints sent to printk as well as the
4041 tracing ring buffer. This is useful for early boot up
4042 where the system hangs or reboots and does not give the
4043 option for reading the tracing buffer or performing a
4044 ftrace_dump_on_oops.
4046 To turn off having tracepoints sent to printk,
4047 echo 0 > /proc/sys/kernel/tracepoint_printk
4048 Note, echoing 1 into this file without the
4049 tracepoint_printk kernel cmdline option has no effect.
4053 Having tracepoints sent to printk() and activating high
4054 frequency tracepoints such as irq or sched, can cause
4055 the system to live lock.
4058 [FTRACE] enable this option to disable tracing when a
4059 warning is hit. This turns off "tracing_on". Tracing can
4060 be enabled again by echoing '1' into the "tracing_on"
4061 file located in /sys/kernel/debug/tracing/
4063 This option is useful, as it disables the trace before
4064 the WARNING dump is called, which prevents the trace to
4065 be filled with content caused by the warning output.
4067 This option can also be set at run time via the sysctl
4068 option: kernel/traceoff_on_warning
4070 transparent_hugepage=
4072 Format: [always|madvise|never]
4073 Can be used to control the default behavior of the system
4074 with respect to transparent hugepages.
4075 See Documentation/vm/transhuge.txt for more details.
4077 tsc= Disable clocksource stability checks for TSC.
4079 [x86] reliable: mark tsc clocksource as reliable, this
4080 disables clocksource verification at runtime, as well
4081 as the stability checks done at bootup. Used to enable
4082 high-resolution timer mode on older hardware, and in
4083 virtualized environment.
4084 [x86] noirqtime: Do not use TSC to do irq accounting.
4085 Used to run time disable IRQ_TIME_ACCOUNTING on any
4086 platforms where RDTSC is slow and this accounting
4089 tsx= [X86] Control Transactional Synchronization
4090 Extensions (TSX) feature in Intel processors that
4091 support TSX control.
4093 This parameter controls the TSX feature. The options are:
4095 on - Enable TSX on the system. Although there are
4096 mitigations for all known security vulnerabilities,
4097 TSX has been known to be an accelerator for
4098 several previous speculation-related CVEs, and
4099 so there may be unknown security risks associated
4100 with leaving it enabled.
4102 off - Disable TSX on the system. (Note that this
4103 option takes effect only on newer CPUs which are
4104 not vulnerable to MDS, i.e., have
4105 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
4106 the new IA32_TSX_CTRL MSR through a microcode
4107 update. This new MSR allows for the reliable
4108 deactivation of the TSX functionality.)
4110 auto - Disable TSX if X86_BUG_TAA is present,
4111 otherwise enable TSX on the system.
4113 Not specifying this option is equivalent to tsx=off.
4115 See Documentation/hw-vuln/tsx_async_abort.rst
4118 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
4119 Abort (TAA) vulnerability.
4121 Similar to Micro-architectural Data Sampling (MDS)
4122 certain CPUs that support Transactional
4123 Synchronization Extensions (TSX) are vulnerable to an
4124 exploit against CPU internal buffers which can forward
4125 information to a disclosure gadget under certain
4128 In vulnerable processors, the speculatively forwarded
4129 data can be used in a cache side channel attack, to
4130 access data to which the attacker does not have direct
4133 This parameter controls the TAA mitigation. The
4136 full - Enable TAA mitigation on vulnerable CPUs
4139 off - Unconditionally disable TAA mitigation
4141 On MDS-affected machines, tsx_async_abort=off can be
4142 prevented by an active MDS mitigation as both vulnerabilities
4143 are mitigated with the same mechanism so in order to disable
4144 this mitigation, you need to specify mds=off too.
4146 Not specifying this option is equivalent to
4147 tsx_async_abort=full. On CPUs which are MDS affected
4148 and deploy MDS mitigation, TAA mitigation is not
4149 required and doesn't provide any additional
4153 Documentation/hw-vuln/tsx_async_abort.rst
4155 turbografx.map[2|3]= [HW,JOY]
4156 TurboGraFX parallel port interface
4158 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
4159 See also Documentation/input/joystick-parport.txt
4161 udbg-immortal [PPC] When debugging early kernel crashes that
4162 happen after console_init() and before a proper
4163 console driver takes over, this boot options might
4164 help "seeing" what's going on.
4166 uhash_entries= [KNL,NET]
4167 Set number of hash buckets for UDP/UDP-Lite connections
4170 [USB] Ignore overcurrent events (default N).
4171 Some badly-designed motherboards generate lots of
4172 bogus events, for ports that aren't wired to
4173 anything. Set this parameter to avoid log spamming.
4174 Note that genuine overcurrent events won't be
4178 [X86] Cause panic on unknown NMI.
4180 usbcore.authorized_default=
4181 [USB] Default USB device authorization:
4182 (default -1 = authorized except for wireless USB,
4183 0 = not authorized, 1 = authorized)
4185 usbcore.autosuspend=
4186 [USB] The autosuspend time delay (in seconds) used
4187 for newly-detected USB devices (default 2). This
4188 is the time required before an idle device will be
4189 autosuspended. Devices for which the delay is set
4190 to a negative value won't be autosuspended at all.
4192 usbcore.usbfs_snoop=
4193 [USB] Set to log all usbfs traffic (default 0 = off).
4195 usbcore.blinkenlights=
4196 [USB] Set to cycle leds on hubs (default 0 = off).
4198 usbcore.old_scheme_first=
4199 [USB] Start with the old device initialization
4200 scheme (default 0 = off).
4202 usbcore.usbfs_memory_mb=
4203 [USB] Memory limit (in MB) for buffers allocated by
4204 usbfs (default = 16, 0 = max = 2047).
4206 usbcore.use_both_schemes=
4207 [USB] Try the other device initialization scheme
4208 if the first one fails (default 1 = enabled).
4210 usbcore.initial_descriptor_timeout=
4211 [USB] Specifies timeout for the initial 64-byte
4212 USB_REQ_GET_DESCRIPTOR request in milliseconds
4213 (default 5000 = 5.0 seconds).
4216 [USBHID] The interval which mice are to be polled at.
4218 usb-storage.delay_use=
4219 [UMS] The delay in seconds before a new device is
4220 scanned for Logical Units (default 1).
4223 [UMS] A list of quirks entries to supplement or
4224 override the built-in unusual_devs list. List
4225 entries are separated by commas. Each entry has
4226 the form VID:PID:Flags where VID and PID are Vendor
4227 and Product ID values (4-digit hex numbers) and
4228 Flags is a set of characters, each corresponding
4229 to a common usb-storage quirk flag as follows:
4230 a = SANE_SENSE (collect more than 18 bytes
4232 b = BAD_SENSE (don't collect more than 18
4233 bytes of sense data);
4234 c = FIX_CAPACITY (decrease the reported
4235 device capacity by one sector);
4236 d = NO_READ_DISC_INFO (don't use
4237 READ_DISC_INFO command);
4238 e = NO_READ_CAPACITY_16 (don't use
4239 READ_CAPACITY_16 command);
4240 f = NO_REPORT_OPCODES (don't use report opcodes
4242 g = MAX_SECTORS_240 (don't transfer more than
4243 240 sectors at a time, uas only);
4244 h = CAPACITY_HEURISTICS (decrease the
4245 reported device capacity by one
4246 sector if the number is odd);
4247 i = IGNORE_DEVICE (don't bind to this
4249 j = NO_REPORT_LUNS (don't use report luns
4251 l = NOT_LOCKABLE (don't try to lock and
4252 unlock ejectable media);
4253 m = MAX_SECTORS_64 (don't transfer more
4254 than 64 sectors = 32 KB at a time);
4255 n = INITIAL_READ10 (force a retry of the
4256 initial READ(10) command);
4257 o = CAPACITY_OK (accept the capacity
4258 reported by the device);
4259 p = WRITE_CACHE (the device cache is ON
4261 r = IGNORE_RESIDUE (the device reports
4262 bogus residue values);
4263 s = SINGLE_LUN (the device has only one
4265 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
4266 commands, uas only);
4267 u = IGNORE_UAS (don't bind to the uas driver);
4268 w = NO_WP_DETECT (don't test whether the
4269 medium is write-protected).
4270 Example: quirks=0419:aaf5:rl,0421:0433:rc
4272 user_debug= [KNL,ARM]
4274 See arch/arm/Kconfig.debug help text.
4275 1 - undefined instruction events
4277 4 - invalid data aborts
4280 Example: user_debug=31
4283 [X86] Flags controlling user PTE allocations.
4285 nohigh = do not allocate PTE pages in
4286 HIGHMEM regardless of setting
4290 On X86_32, this is an alias for vdso32=. Otherwise:
4292 vdso=1: enable VDSO (the default)
4293 vdso=0: disable VDSO mapping
4295 vdso32= [X86] Control the 32-bit vDSO
4296 vdso32=1: enable 32-bit VDSO
4297 vdso32=0 or vdso32=2: disable 32-bit VDSO
4299 See the help text for CONFIG_COMPAT_VDSO for more
4300 details. If CONFIG_COMPAT_VDSO is set, the default is
4301 vdso32=0; otherwise, the default is vdso32=1.
4303 For compatibility with older kernels, vdso32=2 is an
4306 Try vdso32=0 if you encounter an error that says:
4307 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
4310 vector=percpu: enable percpu vector domain
4312 video= [FB] Frame buffer configuration
4313 See Documentation/fb/modedb.txt.
4315 video.brightness_switch_enabled= [0,1]
4316 If set to 1, on receiving an ACPI notify event
4317 generated by hotkey, video driver will adjust brightness
4318 level and then send out the event to user space through
4319 the allocated input device; If set to 0, video driver
4320 will only send out the event without touching backlight
4325 [VMMIO] Memory mapped virtio (platform) device.
4327 <size>@<baseaddr>:<irq>[:<id>]
4329 <size> := size (can use standard suffixes
4331 <baseaddr> := physical base address
4332 <irq> := interrupt number (as passed to
4334 <id> := (optional) platform device id
4336 virtio_mmio.device=1K@0x100b0000:48:7
4338 Can be used multiple times for multiple devices.
4340 vga= [BOOT,X86-32] Select a particular video mode
4341 See Documentation/x86/boot.txt and
4342 Documentation/svga.txt.
4343 Use vga=ask for menu.
4344 This is actually a boot loader parameter; the value is
4345 passed to the kernel using a special protocol.
4347 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
4348 size of <nn>. This can be used to increase the
4349 minimum size (128MB on x86). It can also be used to
4350 decrease the size and leave more room for directly
4353 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
4356 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
4359 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
4363 Controls the behavior of vsyscalls (i.e. calls to
4364 fixed addresses of 0xffffffffff600x00 from legacy
4365 code). Most statically-linked binaries and older
4366 versions of glibc use these calls. Because these
4367 functions are at fixed addresses, they make nice
4368 targets for exploits that can control RIP.
4370 emulate [default] Vsyscalls turn into traps and are
4371 emulated reasonably safely.
4373 native Vsyscalls are native syscall instructions.
4374 This is a little bit faster than trapping
4375 and makes a few dynamic recompilers work
4376 better than they would in emulation mode.
4377 It also makes exploits much easier to write.
4379 none Vsyscalls don't work at all. This makes
4380 them quite hard to use for exploits but
4381 might break your system.
4383 vt.color= [VT] Default text color.
4384 Format: 0xYX, X = foreground, Y = background.
4385 Default: 0x07 = light gray on black.
4387 vt.cur_default= [VT] Default cursor shape.
4388 Format: 0xCCBBAA, where AA, BB, and CC are the same as
4389 the parameters of the <Esc>[?A;B;Cc escape sequence;
4390 see VGA-softcursor.txt. Default: 2 = underline.
4392 vt.default_blu= [VT]
4393 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
4394 Change the default blue palette of the console.
4395 This is a 16-member array composed of values
4398 vt.default_grn= [VT]
4399 Format: <green0>,<green1>,<green2>,...,<green15>
4400 Change the default green palette of the console.
4401 This is a 16-member array composed of values
4404 vt.default_red= [VT]
4405 Format: <red0>,<red1>,<red2>,...,<red15>
4406 Change the default red palette of the console.
4407 This is a 16-member array composed of values
4413 Set system-wide default UTF-8 mode for all tty's.
4414 Default is 1, i.e. UTF-8 mode is enabled for all
4415 newly opened terminals.
4417 vt.global_cursor_default=
4420 Set system-wide default for whether a cursor
4421 is shown on new VTs. Default is -1,
4422 i.e. cursors will be created by default unless
4423 overridden by individual drivers. 0 will hide
4424 cursors, 1 will display them.
4426 vt.italic= [VT] Default color for italic text; 0-15.
4429 vt.underline= [VT] Default color for underlined text; 0-15.
4432 watchdog timers [HW,WDT] For information on watchdog timers,
4433 see Documentation/watchdog/watchdog-parameters.txt
4434 or other driver-specific files in the
4435 Documentation/watchdog/ directory.
4437 workqueue.disable_numa
4438 By default, all work items queued to unbound
4439 workqueues are affine to the NUMA nodes they're
4440 issued on, which results in better behavior in
4441 general. If NUMA affinity needs to be disabled for
4442 whatever reason, this option can be used. Note
4443 that this also can be controlled per-workqueue for
4444 workqueues visible under /sys/bus/workqueue/.
4446 workqueue.power_efficient
4447 Per-cpu workqueues are generally preferred because
4448 they show better performance thanks to cache
4449 locality; unfortunately, per-cpu workqueues tend to
4450 be more power hungry than unbound workqueues.
4452 Enabling this makes the per-cpu workqueues which
4453 were observed to contribute significantly to power
4454 consumption unbound, leading to measurably lower
4455 power usage at the cost of small performance
4458 The default value of this parameter is determined by
4459 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
4461 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
4462 default x2apic cluster mode on platforms
4465 x86_intel_mid_timer= [X86-32,APBT]
4466 Choose timer option for x86 Intel MID platform.
4467 Two valid options are apbt timer only and lapic timer
4468 plus one apbt timer for broadcast timer.
4469 x86_intel_mid_timer=apbt_only | lapic_and_apbt
4471 xen_512gb_limit [KNL,X86-64,XEN]
4472 Restricts the kernel running paravirtualized under Xen
4473 to use only up to 512 GB of RAM. The reason to do so is
4474 crash analysis tools and Xen tools for doing domain
4475 save/restore/migration must be enabled to handle larger
4478 xen_emul_unplug= [HW,X86,XEN]
4479 Unplug Xen emulated devices
4480 Format: [unplug0,][unplug1]
4481 ide-disks -- unplug primary master IDE devices
4482 aux-ide-disks -- unplug non-primary-master IDE devices
4483 nics -- unplug network devices
4484 all -- unplug all emulated devices (NICs and IDE disks)
4485 unnecessary -- unplugging emulated devices is
4486 unnecessary even if the host did not respond to
4488 never -- do not unplug even if version check succeeds
4490 xen_nopvspin [X86,XEN]
4491 Disables the ticketlock slowpath using Xen PV
4495 Disables the PV optimizations forcing the HVM guest to
4496 run as generic HVM guest with no PV drivers.
4498 xen.event_eoi_delay= [XEN]
4499 How long to delay EOI handling in case of event
4500 storms (jiffies). Default is 10.
4502 xen.event_loop_timeout= [XEN]
4503 After which time (jiffies) the event handling loop
4504 should start to delay EOI handling. Default is 2.
4506 xirc2ps_cs= [NET,PCMCIA]
4508 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
4510 ______________________________________________________________________
4514 Add more DRM drivers.